[management] Rework Agent Network endpoint identity and settings bootstrap (#7085)

Store the per-account gateway endpoint as {domain, proxy_address} with a
global unique index on the full hostname; dedicated = (domain ==
proxy_address). Bootstrap becomes an explicit POST carrying exactly one
of proxy_address (server allocates an adjective-noun label beneath it)
or endpoint (claimed verbatim, address-first); provider create loses its
bootstrap side effect. PUT is a full replace with every field required —
the immutable identity fields must be echoed unchanged and a mismatch is
rejected with 422. A guarded DELETE releases the endpoint: refused with
412 while providers exist or a proxy is actively serving the endpoint
hostname (matched case-insensitively); re-creating bootstraps fresh. A
self-addressed pin excludes its address from the account's cluster allow
list, and the live mapping update path now addresses the serving proxy
from the synthesized service. Existing rows are migrated on all three
store engines.
This commit is contained in:
Brad Ison
2026-08-10 19:06:55 +02:00
committed by GitHub
parent e8671a811d
commit ebfdf7d7b8
46 changed files with 2392 additions and 688 deletions
@@ -102,8 +102,8 @@ func TestSettingsHandler_GetExposesCollectionToggles(t *testing.T) {
require.NoError(t, f.store.SaveAgentNetworkSettings(context.Background(), &agentNetworkTypes.Settings{
AccountID: testAccountID,
Cluster: "eu.proxy.netbird.io",
Subdomain: "violet",
Domain: "violet.eu.proxy.netbird.io",
ProxyAddress: "eu.proxy.netbird.io",
EnableLogCollection: true,
EnablePromptCollection: true,
RedactPii: false,
@@ -155,12 +155,7 @@ func (h *handler) createProvider(w http.ResponseWriter, r *http.Request) {
provider := types.NewProvider(userAuth.AccountId)
provider.FromAPIRequest(&req)
bootstrapCluster := ""
if req.BootstrapCluster != nil {
bootstrapCluster = *req.BootstrapCluster
}
created, err := h.manager.CreateProvider(r.Context(), userAuth.UserId, provider, bootstrapCluster)
created, err := h.manager.CreateProvider(r.Context(), userAuth.UserId, provider)
if err != nil {
util.WriteError(r.Context(), err, w)
return
@@ -12,13 +12,55 @@ import (
"github.com/netbirdio/netbird/shared/management/http/util"
)
// addSettingsEndpoints registers the Agent Network settings routes. The
// settings row is bootstrapped server-side on first provider create or on the
// first PUT carrying a cluster; GET reads it and PUT applies a partial update
// of the mutable collection toggles (cluster/subdomain stay immutable).
// addSettingsEndpoints registers the Agent Network settings routes. POST
// bootstraps the settings row, assigning the account's immutable endpoint;
// GET reads it (defaults with an empty endpoint before bootstrap); PUT
// carries every field, replacing the mutable collection toggles and rejecting
// any change to the identity fields; DELETE removes the row — guarded so it
// stays a bootstrap-repair operation — releasing the endpoint for a fresh
// bootstrap.
func (h *handler) addSettingsEndpoints(router *mux.Router) {
router.HandleFunc("/agent-network/settings", h.getSettings).Methods("GET", "OPTIONS")
router.HandleFunc("/agent-network/settings", h.createSettings).Methods("POST", "OPTIONS")
router.HandleFunc("/agent-network/settings", h.updateSettings).Methods("PUT", "OPTIONS")
router.HandleFunc("/agent-network/settings", h.deleteSettings).Methods("DELETE", "OPTIONS")
}
// createSettings bootstraps the account's settings row. Exactly one of
// proxy_address (labeled endpoint; the server allocates the label) and
// endpoint (self-addressed, claimed verbatim) must be provided; optional
// collection toggles ride along with defaults for omitted fields.
func (h *handler) createSettings(w http.ResponseWriter, r *http.Request) {
userAuth, err := nbcontext.GetUserAuthFromContext(r.Context())
if err != nil {
util.WriteError(r.Context(), err, w)
return
}
var req api.AgentNetworkSettingsCreateRequest
if err := json.NewDecoder(r.Body).Decode(&req); err != nil {
util.WriteErrorResponse("couldn't parse JSON request", http.StatusBadRequest, w)
return
}
settings := types.DefaultSettings(userAuth.AccountId)
settings.FromAPICreateRequest(&req)
proxyAddress := ""
if req.ProxyAddress != nil {
proxyAddress = *req.ProxyAddress
}
endpoint := ""
if req.Endpoint != nil {
endpoint = *req.Endpoint
}
created, err := h.manager.CreateSettings(r.Context(), userAuth.UserId, settings, proxyAddress, endpoint)
if err != nil {
util.WriteError(r.Context(), err, w)
return
}
util.WriteJSONObject(r.Context(), w, created.ToAPIResponse())
}
// updateSettings replaces the mutable settings fields on the account's row.
@@ -48,6 +90,24 @@ func (h *handler) updateSettings(w http.ResponseWriter, r *http.Request) {
util.WriteJSONObject(r.Context(), w, updated.ToAPIResponse())
}
// deleteSettings removes the account's settings row, releasing the endpoint.
// The manager refuses (412) while providers exist or a proxy is actively
// serving the endpoint; a later POST bootstraps fresh, allocating a new
// endpoint.
func (h *handler) deleteSettings(w http.ResponseWriter, r *http.Request) {
userAuth, err := nbcontext.GetUserAuthFromContext(r.Context())
if err != nil {
util.WriteError(r.Context(), err, w)
return
}
if err := h.manager.DeleteSettings(r.Context(), userAuth.AccountId, userAuth.UserId); err != nil {
util.WriteError(r.Context(), err, w)
return
}
util.WriteJSONObject(r.Context(), w, util.EmptyObject{})
}
// getSettings returns the account's agent-network settings. Accounts that
// haven't been bootstrapped yet read as the defaults with an empty cluster,
// subdomain and endpoint; the manager synthesises that view.
@@ -1,20 +1,25 @@
package handlers
import (
"context"
"encoding/json"
"fmt"
"net/http"
"strings"
"testing"
"time"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
rpproxy "github.com/netbirdio/netbird/management/internals/modules/reverseproxy/proxy"
"github.com/netbirdio/netbird/shared/management/http/api"
)
// TestSettingsHandler_GetUnbootstrappedReturnsDefaults pins the settings-read
// convention shared with the account and DNS settings endpoints: settings
// always read as a JSON object. Before bootstrap that object carries the
// defaults with an empty cluster/subdomain/endpoint (the "not bootstrapped"
// defaults with an empty endpoint/proxy_address (the "not bootstrapped"
// signal) and no timestamps — never a 404 and never the legacy null body.
func TestSettingsHandler_GetUnbootstrappedReturnsDefaults(t *testing.T) {
f := newAgentNetworkHandlerFixture(t)
@@ -27,9 +32,9 @@ func TestSettingsHandler_GetUnbootstrappedReturnsDefaults(t *testing.T) {
var got api.AgentNetworkSettings
require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &got))
assert.Empty(t, got.Cluster, "cluster must be empty until bootstrapped")
assert.Empty(t, got.Subdomain, "subdomain must be empty until bootstrapped")
assert.Empty(t, got.Endpoint, "endpoint must be empty until bootstrapped, not a bare dot")
assert.Empty(t, got.Endpoint, "endpoint must be empty until bootstrapped")
assert.Empty(t, got.ProxyAddress, "proxy address must be empty until bootstrapped")
assert.False(t, got.Dedicated, "an unbootstrapped account has no serving shape")
assert.True(t, got.EnableLogCollection, "defaults must show log collection on, matching bootstrap")
assert.False(t, got.EnablePromptCollection, "defaults must show prompt collection off")
assert.False(t, got.RedactPii, "defaults must show redaction off")
@@ -39,62 +44,149 @@ func TestSettingsHandler_GetUnbootstrappedReturnsDefaults(t *testing.T) {
assert.Nil(t, got.UpdatedAt, "no timestamps before a row exists")
}
// TestSettingsHandler_PutBootstrapsWithCluster covers the settings-first
// bootstrap path: a PUT carrying a cluster on an unbootstrapped account
// creates the row (cluster pinned, subdomain assigned) and applies the
// mutable fields from the same request.
func TestSettingsHandler_PutBootstrapsWithCluster(t *testing.T) {
// TestSettingsHandler_PostBootstrapsLabeled covers the labeled bootstrap
// shape: a POST carrying a proxy_address allocates a label beneath it, so the
// endpoint hangs one label under the shared cluster's address and the pin is
// not dedicated. Toggles riding along apply; omitted ones keep defaults.
func TestSettingsHandler_PostBootstrapsLabeled(t *testing.T) {
f := newAgentNetworkHandlerFixture(t)
rec := f.do(t, http.MethodPut, "/agent-network/settings",
`{"cluster": "eu.proxy.netbird.io", "enable_log_collection": true, "enable_prompt_collection": true, "redact_pii": false, "access_log_retention_days": 30}`)
require.Equal(t, http.StatusOK, rec.Code, "bootstrap PUT must succeed: %s", rec.Body.String())
rec := f.do(t, http.MethodPost, "/agent-network/settings",
`{"proxy_address": "eu.proxy.netbird.io", "enable_prompt_collection": true, "access_log_retention_days": 14}`)
require.Equal(t, http.StatusOK, rec.Code, "bootstrap POST must succeed: %s", rec.Body.String())
var got api.AgentNetworkSettings
require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &got))
assert.Equal(t, "eu.proxy.netbird.io", got.Cluster, "cluster must be pinned from the request")
assert.NotEmpty(t, got.Subdomain, "subdomain must be assigned at bootstrap")
assert.Equal(t, got.Subdomain+".eu.proxy.netbird.io", got.Endpoint, "endpoint must combine subdomain and cluster")
assert.True(t, got.EnableLogCollection, "toggle from the bootstrap request must apply")
assert.Equal(t, "eu.proxy.netbird.io", got.ProxyAddress, "proxy address must be pinned from the request")
require.NotEmpty(t, got.Endpoint, "endpoint must be allocated at bootstrap")
assert.True(t, strings.HasSuffix(got.Endpoint, ".eu.proxy.netbird.io"),
"labeled endpoint must hang off the proxy address: %s", got.Endpoint)
label := strings.TrimSuffix(got.Endpoint, ".eu.proxy.netbird.io")
assert.NotContains(t, label, ".", "the allocated label must be a single DNS label: %s", label)
assert.False(t, got.Dedicated, "a labeled pin is not dedicated")
assert.True(t, got.EnableLogCollection, "omitted toggle must keep its default")
assert.True(t, got.EnablePromptCollection, "toggle from the bootstrap request must apply")
require.NotNil(t, got.AccessLogRetentionDays)
assert.Equal(t, 30, *got.AccessLogRetentionDays, "retention from the bootstrap request must apply")
assert.Equal(t, 14, *got.AccessLogRetentionDays, "retention from the bootstrap request must apply")
assert.NotNil(t, got.CreatedAt, "a persisted row carries timestamps")
// The row is now readable via GET.
rec = f.do(t, http.MethodGet, "/agent-network/settings", "")
require.Equal(t, http.StatusOK, rec.Code, "GET after bootstrap must succeed")
var read api.AgentNetworkSettings
require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &read))
assert.Equal(t, got.Endpoint, read.Endpoint, "GET must return the bootstrapped endpoint")
}
// TestSettingsHandler_PutWithoutClusterOnUnbootstrapped pins that a PUT
// without a cluster cannot conjure a settings row out of nothing — there is
// no cluster to pin — and surfaces as 404 like the GET.
func TestSettingsHandler_PutWithoutClusterOnUnbootstrapped(t *testing.T) {
// TestSettingsHandler_PostBootstrapsSelfAddressed covers the dedicated shape:
// a POST carrying an endpoint claims the hostname verbatim, the proxy address
// equals it, and the pin reads as dedicated. The claim is legitimate before
// any proxy declares the address (address-first).
func TestSettingsHandler_PostBootstrapsSelfAddressed(t *testing.T) {
f := newAgentNetworkHandlerFixture(t)
rec := f.do(t, http.MethodPost, "/agent-network/settings",
`{"endpoint": "Brave-Otter.Gateway.Example.com"}`)
require.Equal(t, http.StatusOK, rec.Code, "bootstrap POST must succeed: %s", rec.Body.String())
var got api.AgentNetworkSettings
require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &got))
assert.Equal(t, "brave-otter.gateway.example.com", got.Endpoint,
"endpoint must be claimed verbatim, lowercased")
assert.Equal(t, got.Endpoint, got.ProxyAddress, "self-addressed: the proxy address is the endpoint")
assert.True(t, got.Dedicated, "a self-addressed pin is dedicated")
assert.True(t, got.EnableLogCollection, "omitted toggles must keep their defaults")
}
// TestSettingsHandler_PostRequiresExactlyOneIdentityField pins the request
// contract: proxy_address and endpoint are mutually exclusive and one is
// required — both or neither is a validation error, not a guess.
func TestSettingsHandler_PostRequiresExactlyOneIdentityField(t *testing.T) {
f := newAgentNetworkHandlerFixture(t)
rec := f.do(t, http.MethodPost, "/agent-network/settings", `{}`)
assert.Equal(t, http.StatusUnprocessableEntity, rec.Code,
"empty POST must be rejected: got %d body=%s", rec.Code, rec.Body.String())
rec = f.do(t, http.MethodPost, "/agent-network/settings",
`{"proxy_address": "eu.proxy.netbird.io", "endpoint": "brave-otter.gateway.example.com"}`)
assert.Equal(t, http.StatusUnprocessableEntity, rec.Code,
"POST with both identity fields must be rejected: got %d body=%s", rec.Code, rec.Body.String())
}
// TestSettingsHandler_PostRejectsMalformedHostnames pins per-write input
// validation: shapes canonicalization cannot repair — trailing dots, embedded
// whitespace, empty labels — are rejected with a validation error instead of
// landing in an immutable column.
func TestSettingsHandler_PostRejectsMalformedHostnames(t *testing.T) {
f := newAgentNetworkHandlerFixture(t)
for name, body := range map[string]string{
"trailing dot": `{"endpoint": "gateway.example.com."}`,
"leading dot": `{"endpoint": ".gateway.example.com"}`,
"inner whitespace": `{"endpoint": "gate way.example.com"}`,
"empty label": `{"proxy_address": "eu..proxy.netbird.io"}`,
} {
rec := f.do(t, http.MethodPost, "/agent-network/settings", body)
assert.Equal(t, http.StatusUnprocessableEntity, rec.Code,
"%s must be rejected: got %d body=%s", name, rec.Code, rec.Body.String())
}
}
// TestSettingsHandler_PostConflictsOnSecondBootstrap pins that bootstrap is a
// one-time create: a second POST returns 409 and leaves the row untouched.
func TestSettingsHandler_PostConflictsOnSecondBootstrap(t *testing.T) {
f := newAgentNetworkHandlerFixture(t)
rec := f.do(t, http.MethodPost, "/agent-network/settings", `{"proxy_address": "eu.proxy.netbird.io"}`)
require.Equal(t, http.StatusOK, rec.Code, "first bootstrap must succeed: %s", rec.Body.String())
var first api.AgentNetworkSettings
require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &first))
rec = f.do(t, http.MethodPost, "/agent-network/settings", `{"proxy_address": "us.proxy.netbird.io"}`)
assert.Equal(t, http.StatusConflict, rec.Code,
"second bootstrap must 409: got %d body=%s", rec.Code, rec.Body.String())
rec = f.do(t, http.MethodGet, "/agent-network/settings", "")
require.Equal(t, http.StatusOK, rec.Code)
var got api.AgentNetworkSettings
require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &got))
assert.Equal(t, first.Endpoint, got.Endpoint, "the original endpoint must survive the rejected bootstrap")
assert.Equal(t, first.ProxyAddress, got.ProxyAddress, "the original proxy address must survive")
}
// TestSettingsHandler_PutBeforeBootstrapIs404 pins that a PUT cannot conjure a
// settings row out of nothing — bootstrap is the explicit POST — and the
// error points the caller there.
func TestSettingsHandler_PutBeforeBootstrapIs404(t *testing.T) {
f := newAgentNetworkHandlerFixture(t)
rec := f.do(t, http.MethodPut, "/agent-network/settings",
`{"enable_log_collection": false, "enable_prompt_collection": false, "redact_pii": false}`)
assert.Equal(t, http.StatusNotFound, rec.Code,
"cluster-less PUT on an unbootstrapped account must 404: got %d body=%s", rec.Code, rec.Body.String())
assert.Contains(t, rec.Body.String(), "cluster",
"the error must point the caller at the bootstrap paths: %s", rec.Body.String())
"PUT on an unbootstrapped account must 404: got %d body=%s", rec.Code, rec.Body.String())
assert.Contains(t, rec.Body.String(), "/api/agent-network/settings",
"the error must point the caller at the bootstrap POST: %s", rec.Body.String())
}
// TestSettingsHandler_PutReplacesMutableFields pins the update contract shared
// with the other PUT endpoints: the request replaces every mutable field, so a
// toggle absent from the JSON lands as its zero value rather than being
// preserved. Cluster and subdomain survive untouched.
// with the other PUT endpoints: the request carries every field, replacing the
// mutable ones. The identity fields ride along as a required echo of the
// assigned values — compared, never written — so the endpoint and proxy
// address survive every accepted update.
func TestSettingsHandler_PutReplacesMutableFields(t *testing.T) {
f := newAgentNetworkHandlerFixture(t)
rec := f.do(t, http.MethodPut, "/agent-network/settings",
`{"cluster": "eu.proxy.netbird.io", "enable_log_collection": true, "enable_prompt_collection": true, "redact_pii": true, "access_log_retention_days": 14}`)
require.Equal(t, http.StatusOK, rec.Code, "bootstrap PUT must succeed: %s", rec.Body.String())
rec := f.do(t, http.MethodPost, "/agent-network/settings",
`{"proxy_address": "eu.proxy.netbird.io", "enable_prompt_collection": true, "redact_pii": true, "access_log_retention_days": 14}`)
require.Equal(t, http.StatusOK, rec.Code, "bootstrap POST must succeed: %s", rec.Body.String())
var before api.AgentNetworkSettings
require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &before))
rec = f.do(t, http.MethodPut, "/agent-network/settings",
`{"enable_log_collection": true, "enable_prompt_collection": false, "redact_pii": false}`)
rec = f.do(t, http.MethodPut, "/agent-network/settings", fmt.Sprintf(
`{"endpoint": %q, "proxy_address": %q, "enable_log_collection": true, "enable_prompt_collection": false, "redact_pii": false, "access_log_retention_days": 7}`,
before.Endpoint, before.ProxyAddress))
require.Equal(t, http.StatusOK, rec.Code, "update PUT must succeed: %s", rec.Body.String())
var got api.AgentNetworkSettings
@@ -103,35 +195,201 @@ func TestSettingsHandler_PutReplacesMutableFields(t *testing.T) {
assert.False(t, got.EnablePromptCollection, "sent toggle must apply")
assert.False(t, got.RedactPii, "sent toggle must apply")
require.NotNil(t, got.AccessLogRetentionDays)
assert.Equal(t, 0, *got.AccessLogRetentionDays,
"retention absent from the request must land as the zero value — PUT replaces all mutable fields")
assert.Equal(t, before.Cluster, got.Cluster, "cluster must survive updates untouched")
assert.Equal(t, before.Subdomain, got.Subdomain, "subdomain must survive updates untouched")
assert.Equal(t, 7, *got.AccessLogRetentionDays, "sent retention must apply")
assert.Equal(t, before.Endpoint, got.Endpoint, "endpoint must survive updates untouched")
assert.Equal(t, before.ProxyAddress, got.ProxyAddress, "proxy address must survive updates untouched")
}
// TestSettingsHandler_PutRejectsClusterChange pins cluster immutability: once
// assigned, a differing cluster is rejected as a validation error instead of
// being silently ignored, so callers never observe a value other than the one
// they sent. Echoing the assigned cluster back stays valid, which lets
// declarative clients send their full desired state idempotently.
func TestSettingsHandler_PutRejectsClusterChange(t *testing.T) {
// TestSettingsHandler_PutRejectsChangedIdentity pins the immutability contract:
// the PUT carries the identity fields like every other field, but they are an
// echo — a request carrying a different endpoint or proxy address is rejected
// as a validation error and the row is left untouched. The comparison is
// lenient about casing (the stored values are normalized lowercase), so a
// client replaying a GET response with different casing is not rejected.
func TestSettingsHandler_PutRejectsChangedIdentity(t *testing.T) {
f := newAgentNetworkHandlerFixture(t)
rec := f.do(t, http.MethodPut, "/agent-network/settings",
`{"cluster": "eu.proxy.netbird.io", "enable_log_collection": true, "enable_prompt_collection": false, "redact_pii": false}`)
require.Equal(t, http.StatusOK, rec.Code, "bootstrap PUT must succeed: %s", rec.Body.String())
rec := f.do(t, http.MethodPost, "/agent-network/settings",
`{"proxy_address": "eu.proxy.netbird.io", "enable_prompt_collection": true}`)
require.Equal(t, http.StatusOK, rec.Code, "bootstrap POST must succeed: %s", rec.Body.String())
var before api.AgentNetworkSettings
require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &before))
rec = f.do(t, http.MethodPut, "/agent-network/settings",
`{"cluster": "us.proxy.netbird.io", "enable_log_collection": true, "enable_prompt_collection": false, "redact_pii": false}`)
assert.Equal(t, http.StatusUnprocessableEntity, rec.Code,
"cluster change must be rejected as a validation error: got %d body=%s", rec.Code, rec.Body.String())
for name, body := range map[string]string{
"changed endpoint": fmt.Sprintf(
`{"endpoint": "other.gateway.example.com", "proxy_address": %q, "enable_log_collection": false, "enable_prompt_collection": false, "redact_pii": false, "access_log_retention_days": 30}`,
before.ProxyAddress),
"changed proxy_address": fmt.Sprintf(
`{"endpoint": %q, "proxy_address": "us.proxy.netbird.io", "enable_log_collection": false, "enable_prompt_collection": false, "redact_pii": false, "access_log_retention_days": 30}`,
before.Endpoint),
"omitted identity": `{"enable_log_collection": false, "enable_prompt_collection": false, "redact_pii": false, "access_log_retention_days": 30}`,
} {
rec = f.do(t, http.MethodPut, "/agent-network/settings", body)
assert.Equal(t, http.StatusUnprocessableEntity, rec.Code,
"%s must be rejected: got %d body=%s", name, rec.Code, rec.Body.String())
}
rec = f.do(t, http.MethodPut, "/agent-network/settings",
`{"cluster": "eu.proxy.netbird.io", "enable_log_collection": true, "enable_prompt_collection": false, "redact_pii": true}`)
require.Equal(t, http.StatusOK, rec.Code, "echoing the assigned cluster must stay valid: %s", rec.Body.String())
// The rejected updates must not have applied anything — toggles included.
rec = f.do(t, http.MethodGet, "/agent-network/settings", "")
require.Equal(t, http.StatusOK, rec.Code)
var got api.AgentNetworkSettings
require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &got))
assert.Equal(t, before.Endpoint, got.Endpoint, "rejected PUT must not change the endpoint")
assert.Equal(t, before.ProxyAddress, got.ProxyAddress, "rejected PUT must not change the proxy address")
assert.True(t, got.EnablePromptCollection, "rejected PUT must not apply its toggles")
// An uppercased echo of the assigned values still names the same host and
// must be accepted.
rec = f.do(t, http.MethodPut, "/agent-network/settings", fmt.Sprintf(
`{"endpoint": %q, "proxy_address": %q, "enable_log_collection": true, "enable_prompt_collection": true, "redact_pii": false, "access_log_retention_days": 30}`,
strings.ToUpper(before.Endpoint), strings.ToUpper(before.ProxyAddress)))
assert.Equal(t, http.StatusOK, rec.Code,
"an uppercased identity echo must be accepted: got %d body=%s", rec.Code, rec.Body.String())
}
// TestSettingsHandler_PutOmittedRetentionLandsAsZero documents a residual the
// required-ness of access_log_retention_days does not remove. Marking the field
// required changes the generated client type from *int to int, so a generated
// client cannot omit it — but nothing validates OpenAPI required-ness at
// runtime, so a hand-rolled body without the field still decodes as 0, which
// the API documents as "keep indefinitely".
//
// That is the same latitude the three booleans already have, so it is left
// consistent rather than special-cased. This test exists to make the gap
// explicit: if request validation is ever added, this expectation is what
// changes.
func TestSettingsHandler_PutOmittedRetentionLandsAsZero(t *testing.T) {
f := newAgentNetworkHandlerFixture(t)
rec := f.do(t, http.MethodPost, "/agent-network/settings",
`{"proxy_address": "eu.proxy.netbird.io", "access_log_retention_days": 14}`)
require.Equal(t, http.StatusOK, rec.Code, "bootstrap POST must succeed: %s", rec.Body.String())
var before api.AgentNetworkSettings
require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &before))
rec = f.do(t, http.MethodPut, "/agent-network/settings", fmt.Sprintf(
`{"endpoint": %q, "proxy_address": %q, "enable_log_collection": true, "enable_prompt_collection": false, "redact_pii": false}`,
before.Endpoint, before.ProxyAddress))
require.Equal(t, http.StatusOK, rec.Code, "update PUT must succeed: %s", rec.Body.String())
var got api.AgentNetworkSettings
require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &got))
assert.Equal(t, "eu.proxy.netbird.io", got.Cluster, "cluster must be unchanged")
assert.True(t, got.RedactPii, "toggle sent alongside the echoed cluster must apply")
require.NotNil(t, got.AccessLogRetentionDays)
assert.Equal(t, 0, *got.AccessLogRetentionDays,
"a non-conforming body that omits retention still replaces it with the zero value")
}
// TestSettingsHandler_DeleteBeforeBootstrapIs404 pins that DELETE on an
// account with no settings row is a 404, mirroring the PUT.
func TestSettingsHandler_DeleteBeforeBootstrapIs404(t *testing.T) {
f := newAgentNetworkHandlerFixture(t)
rec := f.do(t, http.MethodDelete, "/agent-network/settings", "")
assert.Equal(t, http.StatusNotFound, rec.Code,
"DELETE on an unbootstrapped account must 404: got %d body=%s", rec.Code, rec.Body.String())
}
// TestSettingsHandler_DeleteBlockedByProviders pins the first delete guard:
// while any provider exists for the account, the delete is refused with 412
// and the row survives. Providers route through the endpoint — the guard
// keeps DELETE a bootstrap-repair operation rather than a way to abandon a
// configured gateway.
func TestSettingsHandler_DeleteBlockedByProviders(t *testing.T) {
f := newAgentNetworkHandlerFixture(t)
rec := f.do(t, http.MethodPost, "/agent-network/settings", `{"proxy_address": "eu.proxy.netbird.io"}`)
require.Equal(t, http.StatusOK, rec.Code, "bootstrap POST must succeed: %s", rec.Body.String())
var before api.AgentNetworkSettings
require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &before))
f.seedProvider(t, "prov-guard")
rec = f.do(t, http.MethodDelete, "/agent-network/settings", "")
assert.Equal(t, http.StatusPreconditionFailed, rec.Code,
"delete with a provider present must be refused: got %d body=%s", rec.Code, rec.Body.String())
rec = f.do(t, http.MethodGet, "/agent-network/settings", "")
require.Equal(t, http.StatusOK, rec.Code)
var got api.AgentNetworkSettings
require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &got))
assert.Equal(t, before.Endpoint, got.Endpoint, "the refused delete must leave the row intact")
}
// TestSettingsHandler_DeleteBlockedByActiveProxy pins the second delete
// guard: while a proxy is actively serving the endpoint — an active proxy
// row declaring the endpoint hostname as its cluster address, the dedicated
// shape — the delete is refused with 412. A proxy that has disconnected no
// longer blocks: the guard is about a live serving path, not history.
//
// The proxy declares its address with mixed casing on purpose: Connect
// stores the declared address verbatim while the settings row is normalized
// lowercase, and hostnames are case-insensitive, so the guard must match
// across the casing difference rather than be sidestepped by it.
func TestSettingsHandler_DeleteBlockedByActiveProxy(t *testing.T) {
f := newAgentNetworkHandlerFixture(t)
const endpoint = "gw.dedicated.example.com"
rec := f.do(t, http.MethodPost, "/agent-network/settings", fmt.Sprintf(`{"endpoint": %q}`, endpoint))
require.Equal(t, http.StatusOK, rec.Code, "bootstrap POST must succeed: %s", rec.Body.String())
now := time.Now()
accountID := testAccountID
proxyRow := &rpproxy.Proxy{
ID: "proxy-guard",
SessionID: "sess-1",
ClusterAddress: "GW.Dedicated.Example.Com",
AccountID: &accountID,
LastSeen: now,
ConnectedAt: &now,
Status: rpproxy.StatusConnected,
}
require.NoError(t, f.store.SaveProxy(context.Background(), proxyRow))
rec = f.do(t, http.MethodDelete, "/agent-network/settings", "")
assert.Equal(t, http.StatusPreconditionFailed, rec.Code,
"delete with an active proxy at the endpoint must be refused: got %d body=%s", rec.Code, rec.Body.String())
// Once the proxy disconnects it no longer serves the endpoint, so the
// delete goes through.
require.NoError(t, f.store.DisconnectProxy(context.Background(), proxyRow.ID, proxyRow.SessionID))
rec = f.do(t, http.MethodDelete, "/agent-network/settings", "")
assert.Equal(t, http.StatusOK, rec.Code,
"delete after the proxy disconnected must succeed: got %d body=%s", rec.Code, rec.Body.String())
}
// TestSettingsHandler_DeleteReleasesEndpointForFreshBootstrap pins the
// full-reset semantic that gives replace-on-change clients (e.g. Terraform's
// RequiresReplace) a real path: with both guards clear the delete succeeds,
// the account reads as the defaults again, and a fresh bootstrap draws a
// fresh label. The released hostname is not reserved — a fresh draw may even
// legitimately re-pick it — so the assertions check the new row's shape, not
// that the label differs.
func TestSettingsHandler_DeleteReleasesEndpointForFreshBootstrap(t *testing.T) {
f := newAgentNetworkHandlerFixture(t)
rec := f.do(t, http.MethodPost, "/agent-network/settings",
`{"proxy_address": "eu.proxy.netbird.io", "enable_prompt_collection": true}`)
require.Equal(t, http.StatusOK, rec.Code, "bootstrap POST must succeed: %s", rec.Body.String())
rec = f.do(t, http.MethodDelete, "/agent-network/settings", "")
require.Equal(t, http.StatusOK, rec.Code,
"delete with both guards clear must succeed: got %d body=%s", rec.Code, rec.Body.String())
rec = f.do(t, http.MethodGet, "/agent-network/settings", "")
require.Equal(t, http.StatusOK, rec.Code)
var after api.AgentNetworkSettings
require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &after))
assert.Empty(t, after.Endpoint, "a deleted account must read as unbootstrapped defaults")
assert.False(t, after.EnablePromptCollection, "the deleted row's toggles must not linger")
rec = f.do(t, http.MethodPost, "/agent-network/settings", `{"proxy_address": "eu.proxy.netbird.io"}`)
require.Equal(t, http.StatusOK, rec.Code, "re-bootstrap after delete must succeed: %s", rec.Body.String())
var second api.AgentNetworkSettings
require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &second))
require.NotEmpty(t, second.Endpoint, "the fresh bootstrap must allocate an endpoint")
assert.True(t, strings.HasSuffix(second.Endpoint, ".eu.proxy.netbird.io"),
"the fresh endpoint must hang beneath the requested proxy address: %s", second.Endpoint)
assert.False(t, second.EnablePromptCollection,
"the fresh row must carry bootstrap defaults, not the deleted row's toggles")
assert.NotNil(t, second.CreatedAt, "the fresh row is persisted and carries timestamps")
}
@@ -0,0 +1,37 @@
package labelgen
// adjectives is the descriptor half of a generated label. It pairs with the
// noun pool in words.go to form `<adjective>-<noun>` labels, and is kept
// separate because words.go is almost entirely nouns — drawing both halves
// from it produced unreadable pairs like "millet-hammock". Entries are
// lowercase ASCII, 4-12 chars, free of hyphens and digits, screened for
// offensive/brand/region-specific terms, and disjoint from the noun pool
// (enforced by TestAdjectives_AreDisjointFromNouns).
var adjectives = []string{
"able", "active", "adept", "agile", "airy", "alert", "amiable", "ample",
"ancient", "ardent", "artful", "astute", "balmy", "blithe", "bold", "bonny",
"brave", "breezy", "brisk", "bubbly", "buoyant", "bushy", "candid", "canny",
"cheery", "chilly", "chipper", "chunky", "civil", "classic", "clever", "comely",
"compact", "cordial", "cosmic", "courtly", "crafty", "creamy", "crisp", "cuddly",
"curious", "dainty", "dapper", "daring", "dashing", "deft", "dewy", "diligent",
"downy", "dreamy", "dulcet", "durable", "dusky", "eager", "earnest", "earthy",
"easy", "elated", "elegant", "epic", "fabled", "faithful", "fancy", "fearless",
"feisty", "fervent", "fleet", "fluffy", "fond", "frisky", "frosty", "gallant",
"genial", "genteel", "gentle", "giddy", "gilded", "glad", "glassy", "gleaming",
"glossy", "graceful", "grand", "grainy", "hale", "hardy", "hearty", "hefty",
"honest", "hopeful", "humble", "hushed", "immense", "jaunty", "jolly", "jovial",
"joyful", "jubilant", "keen", "kindly", "kindred", "lanky", "leafy", "limber",
"lively", "lofty", "loyal", "lucent", "lucid", "luminous", "lush", "maroon",
"mellow", "merry", "mighty", "mindful", "mirthful", "misty", "modest", "muted",
"nifty", "nimble", "noble", "patient", "peaceful", "pearly", "peppy", "perky",
"petite", "placid", "playful", "pleasant", "plucky", "plush", "polite", "posh",
"prancing", "pristine", "prompt", "proud", "prudent", "quaint", "quick", "quirky",
"radiant", "ready", "regal", "restful", "robust", "rosy", "ruddy", "rugged",
"sandy", "satin", "saucy", "savvy", "sedate", "serene", "shady", "shiny",
"silken", "silky", "sincere", "sleek", "slender", "smart", "smooth", "snappy",
"snug", "soaring", "sparkly", "spiffy", "spirited", "sprightly", "spry", "stalwart",
"stately", "steady", "sterling", "stoic", "stormy", "stout", "sturdy", "sunlit",
"supple", "svelte", "tawny", "tender", "tidy", "timeless", "trusty", "upbeat",
"urbane", "valiant", "vast", "vernal", "vibrant", "vintage", "whimsy", "willing",
"windy", "winsome", "wintry", "witty", "worthy", "zesty", "zippy",
}
@@ -64,3 +64,20 @@ func PickUnique(rng *rand.Rand, taken map[string]struct{}, fallbackSuffix string
w := pool[rng.Intn(len(pool))]
return fmt.Sprintf("%s-%s", w, fallbackSuffix)
}
// PickTuple returns an adjective-noun label such as "brave-otter". It is still
// a single DNS label.
//
// Unlike PickUnique it takes no `taken` set and has no fallback suffix. The
// noun pool holds 857 entries, which is ample per cluster but a hard ceiling
// once labels must be unique across one shared zone; pairing an adjective with
// a noun spans len(adjectives) * 857 instead. Uniqueness is enforced by a
// database constraint and retried by the caller, rather than guessed from a
// pre-read set that a concurrent allocation can invalidate.
func PickTuple(rng *rand.Rand) string {
nouns := uniqueWords()
if len(nouns) == 0 || len(adjectives) == 0 {
return ""
}
return adjectives[rng.Intn(len(adjectives))] + "-" + nouns[rng.Intn(len(nouns))]
}
@@ -99,3 +99,82 @@ func TestUniqueWords_DropsDuplicates(t *testing.T) {
}
assert.GreaterOrEqual(t, len(pool), 500, "Pool must contain at least 500 unique words")
}
// TestPickTuple_ShapeAndPoolMembership locks the wire-visible shape: an
// adjective and a noun, each from its own pool, joined by a single hyphen so
// the result stays one DNS label.
func TestPickTuple_ShapeAndPoolMembership(t *testing.T) {
nouns := uniqueWords()
inNouns := make(map[string]struct{}, len(nouns))
for _, w := range nouns {
inNouns[w] = struct{}{}
}
inAdjectives := make(map[string]struct{}, len(adjectives))
for _, a := range adjectives {
inAdjectives[a] = struct{}{}
}
rng := rand.New(rand.NewSource(7))
for i := 0; i < 200; i++ {
got := PickTuple(rng)
parts := strings.Split(got, "-")
require.Len(t, parts, 2, "PickTuple must produce exactly two hyphen-joined words; got %q", got)
_, adjOK := inAdjectives[parts[0]]
assert.True(t, adjOK, "First half must be an adjective; %q not in adjectives (from %q)", parts[0], got)
_, nounOK := inNouns[parts[1]]
assert.True(t, nounOK, "Second half must be a noun; %q not in words (from %q)", parts[1], got)
assert.LessOrEqual(t, len(got), 63, "Label must fit a DNS label; got %q (%d chars)", got, len(got))
}
}
// TestAdjectives_AreDisjointFromNouns keeps the namespace a clean product and
// prevents nonsense like "azure-azure": a handful of the noun pool's entries
// are adjectival, and any overlap would let the same word land on both sides.
func TestAdjectives_AreDisjointFromNouns(t *testing.T) {
nouns := make(map[string]struct{}, len(uniqueWords()))
for _, w := range uniqueWords() {
nouns[w] = struct{}{}
}
for _, a := range adjectives {
_, clash := nouns[a]
assert.False(t, clash, "Adjective %q also appears in the noun pool; remove it from one list", a)
}
}
// TestAdjectives_AreDNSSafeAndDeduplicated mirrors the curation contract stated
// in words.go: lowercase ASCII, 4-12 chars, no digits or hyphens, no repeats.
func TestAdjectives_AreDNSSafeAndDeduplicated(t *testing.T) {
seen := make(map[string]struct{}, len(adjectives))
for _, a := range adjectives {
_, dup := seen[a]
assert.False(t, dup, "Duplicate adjective %q", a)
seen[a] = struct{}{}
assert.Regexp(t, `^[a-z]{4,12}$`, a, "Adjective %q must be 4-12 lowercase ASCII letters", a)
}
assert.Greater(t, len(adjectives), 150, "Adjective pool too small to give a useful namespace")
}
// TestPickTuple_DeterministicWithSeededRng documents that generation is a pure
// function of the rng, which is what makes allocation retries reproducible in tests.
func TestPickTuple_DeterministicWithSeededRng(t *testing.T) {
a := PickTuple(rand.New(rand.NewSource(42)))
b := PickTuple(rand.New(rand.NewSource(42)))
assert.Equal(t, a, b, "Same seed must yield the same tuple")
}
// TestPickTuple_SpansALargeNamespace guards the reason we moved to tuples: a
// single-word pool caps the GLOBAL namespace at 857. Drawing many tuples must
// yield overwhelmingly distinct values.
func TestPickTuple_SpansALargeNamespace(t *testing.T) {
rng := rand.New(rand.NewSource(11))
seen := make(map[string]struct{}, 2000)
for i := 0; i < 2000; i++ {
seen[PickTuple(rng)] = struct{}{}
}
assert.Greater(t, len(seen), 1900,
"2000 draws should be nearly all distinct across a ~200k namespace; got %d unique", len(seen))
}
@@ -22,7 +22,6 @@ import (
"github.com/netbirdio/netbird/management/server/permissions/modules"
"github.com/netbirdio/netbird/management/server/permissions/operations"
"github.com/netbirdio/netbird/management/server/store"
"github.com/netbirdio/netbird/shared/management/proto"
"github.com/netbirdio/netbird/shared/management/status"
)
@@ -48,7 +47,7 @@ func ensureSessionKeys(p *types.Provider) error {
type Manager interface {
GetAllProviders(ctx context.Context, accountID, userID string) ([]*types.Provider, error)
GetProvider(ctx context.Context, accountID, userID, providerID string) (*types.Provider, error)
CreateProvider(ctx context.Context, userID string, provider *types.Provider, bootstrapCluster string) (*types.Provider, error)
CreateProvider(ctx context.Context, userID string, provider *types.Provider) (*types.Provider, error)
UpdateProvider(ctx context.Context, userID string, provider *types.Provider) (*types.Provider, error)
DeleteProvider(ctx context.Context, accountID, userID, providerID string) error
@@ -71,7 +70,9 @@ type Manager interface {
DeleteBudgetRule(ctx context.Context, accountID, userID, ruleID string) error
GetSettings(ctx context.Context, accountID, userID string) (*types.Settings, error)
CreateSettings(ctx context.Context, userID string, settings *types.Settings, proxyAddress, endpoint string) (*types.Settings, error)
UpdateSettings(ctx context.Context, userID string, settings *types.Settings) (*types.Settings, error)
DeleteSettings(ctx context.Context, accountID, userID string) error
ListConsumption(ctx context.Context, accountID, userID string) ([]*types.Consumption, error)
ListAccessLogs(ctx context.Context, accountID, userID string, filter types.AgentNetworkAccessLogFilter) ([]*types.AgentNetworkAccessLog, int64, error)
@@ -123,11 +124,10 @@ type managerImpl struct {
proxyController proxy.Controller
// reconcileCache holds the last set of synthesised proxy mappings
// per account so reconcile can emit precise Create/Update/Delete
// updates instead of a full re-push on every mutation. Keyed by
// accountID, then by synthesised service ID.
// per account, each paired with the proxy that served it, so a change
// of serving proxy can be diffed without re-deriving it.
reconcileMu sync.Mutex
reconcileCache map[string]map[string]*proto.ProxyMapping
reconcileCache map[string]map[string]syntheticMapping
// labelRngMu guards labelRng. PickUnique consumes math/rand.Source
// state; concurrent provider creates would otherwise race.
@@ -151,7 +151,7 @@ func NewManager(
accountManager: accountManager,
permissionsManager: permissionsManager,
proxyController: proxyController,
reconcileCache: make(map[string]map[string]*proto.ProxyMapping),
reconcileCache: make(map[string]map[string]syntheticMapping),
labelRng: rand.New(rand.NewSource(time.Now().UnixNano())),
}
}
@@ -170,19 +170,14 @@ func (m *managerImpl) GetProvider(ctx context.Context, accountID, userID, provid
return m.store.GetAgentNetworkProviderByID(ctx, store.LockingStrengthNone, accountID, providerID)
}
// CreateProvider persists a new provider for the account. bootstrapCluster
// is used only when the per-account agent-network Settings row hasn't
// been created yet; otherwise it is ignored (the cluster is pinned on
// Settings and every provider in the account routes through it).
func (m *managerImpl) CreateProvider(ctx context.Context, userID string, provider *types.Provider, bootstrapCluster string) (*types.Provider, error) {
// CreateProvider persists a new provider for the account. Providers have no
// settings side effects: the account's endpoint is bootstrapped separately and
// explicitly via CreateSettings, and every provider in the account routes
// through it.
func (m *managerImpl) CreateProvider(ctx context.Context, userID string, provider *types.Provider) (*types.Provider, error) {
if err := m.requirePermission(ctx, provider.AccountID, userID, modules.AgentNetworkProviders, operations.Create); err != nil {
return nil, err
}
if strings.TrimSpace(bootstrapCluster) != "" {
if err := m.requireSettingsBootstrapPermission(ctx, provider.AccountID, userID); err != nil {
return nil, err
}
}
// An empty api_key would silently produce a synthesised service
// that 401s on every upstream request. Surface the misconfiguration
@@ -206,16 +201,6 @@ func (m *managerImpl) CreateProvider(ctx context.Context, userID string, provide
return nil, fmt.Errorf("save agent network provider: %w", err)
}
if strings.TrimSpace(bootstrapCluster) != "" {
if _, err := m.bootstrapSettingsIfNeeded(ctx, m.store, provider.AccountID, bootstrapCluster); err != nil {
// The provider create has already succeeded; logging the
// bootstrap miss matches the plan's PoC behaviour. The synth
// path treats a missing settings row as a no-op, and the next
// provider create retries the bootstrap.
log.WithContext(ctx).Debugf("agent-network bootstrap settings for account %s on cluster %s: %v", provider.AccountID, bootstrapCluster, err)
}
}
m.accountManager.StoreEvent(ctx, userID, provider.ID, provider.AccountID, activity.AgentNetworkProviderCreated, provider.EventMeta())
m.reconcile(ctx, provider.AccountID)
@@ -560,52 +545,44 @@ func (m *managerImpl) DeleteBudgetRule(ctx context.Context, accountID, userID, r
}
// UpdateSettings replaces the mutable account-level settings — the collection
// toggles and retention — on the account's row. When the account has no
// settings row yet, a non-empty settings.Cluster bootstraps one (same path as
// first provider create); without it the update fails with NotFound. On an
// existing row the cluster and subdomain are immutable: a differing
// settings.Cluster is rejected rather than silently ignored so callers never
// observe a value other than what they sent. Because the collection toggles
// change the synthesised service config (prompt-capture gating, access-log
// emission), a reconcile is triggered so the proxy and peer network maps
// converge on the new state.
// toggles and retention — on the account's row. The identity fields (Domain,
// ProxyAddress) are assigned at bootstrap (CreateSettings) and immutable: the
// request carries them, matching the PUT convention of every other endpoint,
// but they are only compared against the stored row — a request carrying
// different values is rejected, and the stored values are never overwritten.
// When the account has no settings row yet the update fails with NotFound.
// Because the collection toggles change the synthesised service config
// (prompt-capture gating, access-log emission), a reconcile is triggered so
// the proxy and peer network maps converge on the new state.
func (m *managerImpl) UpdateSettings(ctx context.Context, userID string, settings *types.Settings) (*types.Settings, error) {
if err := m.requirePermission(ctx, settings.AccountID, userID, modules.AgentNetworkSettings, operations.Update); err != nil {
return nil, err
}
requestedCluster := strings.TrimSpace(settings.Cluster)
// The row lock from LockingStrengthUpdate only holds for the duration of
// the surrounding transaction, so the read, the cluster-immutability
// check, and the save must share one — otherwise concurrent PUTs could
// interleave between them.
// the surrounding transaction, so the read and the save must share one —
// otherwise concurrent PUTs could interleave between them.
var updated *types.Settings
err := m.store.ExecuteInTransaction(ctx, func(tx store.Store) error {
existing, err := tx.GetAgentNetworkSettings(ctx, store.LockingStrengthUpdate, settings.AccountID)
switch {
case err == nil:
if requestedCluster != "" && requestedCluster != existing.Cluster {
return status.Errorf(status.InvalidArgument, "cluster is immutable once assigned (current: %s)", existing.Cluster)
}
case isNotFound(err):
if requestedCluster == "" {
return status.Errorf(status.NotFound, "agent network settings have not been bootstrapped yet; pass cluster to bootstrap them, or create a provider with bootstrap_cluster set")
}
// Bootstrapping pins the cluster and subdomain — a settings
// create on top of the update the caller already passed, matching
// the gate on the provider-create bootstrap path.
if err := m.requirePermission(ctx, settings.AccountID, userID, modules.AgentNetworkSettings, operations.Create); err != nil {
return err
}
existing, err = m.bootstrapSettingsIfNeeded(ctx, tx, settings.AccountID, requestedCluster)
if err != nil {
return err
}
return status.Errorf(status.NotFound, "agent network settings have not been bootstrapped yet; POST /api/agent-network/settings to bootstrap them")
default:
return fmt.Errorf("get agent network settings: %w", err)
}
// The identity echo is compared leniently (trimmed, case-insensitive):
// the stored values are normalized lowercase, and a client replaying a
// GET response must never be rejected over casing it didn't choose.
if !hostnamesEquivalent(settings.Domain, existing.Domain) {
return status.Errorf(status.InvalidArgument, "endpoint is immutable: it must match the assigned endpoint %q; delete the settings to release it and bootstrap again", existing.Domain)
}
if !hostnamesEquivalent(settings.ProxyAddress, existing.ProxyAddress) {
return status.Errorf(status.InvalidArgument, "proxy_address is immutable: it must match the assigned proxy address %q; delete the settings to release it and bootstrap again", existing.ProxyAddress)
}
existing.EnableLogCollection = settings.EnableLogCollection
existing.EnablePromptCollection = settings.EnablePromptCollection
existing.RedactPii = settings.RedactPii
@@ -632,6 +609,83 @@ func (m *managerImpl) UpdateSettings(ctx context.Context, userID string, setting
return updated, nil
}
// hostnamesEquivalent reports whether a caller-supplied hostname names the
// same host as a stored (normalized, lowercase) one: equal after trimming and
// case folding. No structural validation — an arbitrary mismatch and a
// malformed value are both simply "not the assigned value".
func hostnamesEquivalent(supplied, stored string) bool {
return strings.EqualFold(strings.TrimSpace(supplied), stored)
}
// DeleteSettings removes the account's settings row, releasing the endpoint.
// Two guards make this a bootstrap-repair operation rather than a way to tear
// down a serving gateway, both re-checked under the row lock:
//
// - No Agent Network providers may exist for the account. Providers route
// through the endpoint; delete them first.
// - No proxy may be actively serving the endpoint — that is, no active proxy
// declares the endpoint hostname as its cluster address. This is the
// dedicated (self-addressed) shape's guard: the proxy at the address IS
// this account's gateway. A labeled endpoint hangs beneath a shared
// cluster's address, and with the account's providers already gone the
// shared proxy serves nothing of the account's, so the parent cluster
// being up does not block the delete.
//
// Bootstrapping again after a delete allocates fresh — the released hostname
// is not reserved. That full-reset semantic is what gives clients that model
// immutability as replace-on-change (e.g. Terraform's RequiresReplace) a real
// path: tear down providers, delete, re-create.
func (m *managerImpl) DeleteSettings(ctx context.Context, accountID, userID string) error {
if err := m.requirePermission(ctx, accountID, userID, modules.AgentNetworkSettings, operations.Delete); err != nil {
return err
}
var deleted *types.Settings
err := m.store.ExecuteInTransaction(ctx, func(tx store.Store) error {
existing, err := tx.GetAgentNetworkSettings(ctx, store.LockingStrengthUpdate, accountID)
switch {
case err == nil:
case isNotFound(err):
return status.Errorf(status.NotFound, "agent network settings have not been bootstrapped yet; there is nothing to delete")
default:
return fmt.Errorf("get agent network settings: %w", err)
}
providers, err := tx.GetAccountAgentNetworkProviders(ctx, store.LockingStrengthNone, accountID)
if err != nil {
return fmt.Errorf("get agent network providers: %w", err)
}
if len(providers) > 0 {
return status.Errorf(status.PreconditionFailed, "agent network settings cannot be deleted while %d provider(s) exist; delete the providers first", len(providers))
}
serving, err := tx.HasActiveProxyAtClusterAddress(ctx, existing.Domain)
if err != nil {
return fmt.Errorf("check for a proxy serving the endpoint: %w", err)
}
if serving {
return status.Errorf(status.PreconditionFailed, "agent network settings cannot be deleted while a proxy is actively serving the endpoint %q", existing.Domain)
}
if err := tx.DeleteAgentNetworkSettings(ctx, accountID); err != nil {
return fmt.Errorf("delete agent network settings: %w", err)
}
deleted = existing
return nil
})
if err != nil {
return err
}
m.accountManager.StoreEvent(ctx, userID, accountID, accountID, activity.AgentNetworkSettingsDeleted, map[string]any{
"endpoint": deleted.Domain,
"proxy_address": deleted.ProxyAddress,
})
m.reconcile(ctx, accountID)
return nil
}
// isNotFound reports whether err is a status.NotFound error.
func isNotFound(err error) bool {
var sErr *status.Error
@@ -678,74 +732,162 @@ func (m *managerImpl) GetSettings(ctx context.Context, accountID, userID string)
}
}
// requireSettingsBootstrapPermission gates the one-time settings bootstrap a
// first provider create performs. Pinning the account's cluster and subdomain
// is a settings write, so it needs the settings permission on top of the
// provider one. No-op once the settings row exists.
func (m *managerImpl) requireSettingsBootstrapPermission(ctx context.Context, accountID, userID string) error {
_, err := m.store.GetAgentNetworkSettings(ctx, store.LockingStrengthNone, accountID)
if err == nil {
return nil
}
if !isNotFound(err) {
return fmt.Errorf("get agent network settings: %w", err)
}
return m.requirePermission(ctx, accountID, userID, modules.AgentNetworkSettings, operations.Create)
}
// maxDomainAllocationAttempts bounds the label search when bootstrapping a
// labeled endpoint. Package-level (rather than function-local) so tests can
// assert on the exhaustion path without duplicating the literal.
const maxDomainAllocationAttempts = 10
// bootstrapSettingsIfNeeded creates the per-account agent-network
// settings row when missing. The cluster comes from the create-time
// hint the dashboard sends (auto-picked from the active cluster list);
// the subdomain is picked from the curated wordlist avoiding
// collisions on the same cluster. Idempotent: if a row already exists
// it is returned untouched and the hint is ignored. st is the store to
// operate on — pass the transaction store when calling from within one.
func (m *managerImpl) bootstrapSettingsIfNeeded(ctx context.Context, st store.Store, accountID, providerCluster string) (*types.Settings, error) {
if accountID == "" {
return nil, fmt.Errorf("bootstrap settings: account id is required")
// CreateSettings bootstraps the per-account settings row, assigning the
// account's immutable endpoint. Exactly one of proxyAddress and endpoint must
// be non-empty: proxyAddress allocates a labeled endpoint one label beneath
// the given cluster address; endpoint claims the given hostname verbatim as a
// self-addressed (dedicated) endpoint — a legitimate claim before any proxy
// declares the address (address-first). settings carries the account ID and
// the initial collection toggles; its identity fields are assigned here.
func (m *managerImpl) CreateSettings(ctx context.Context, userID string, settings *types.Settings, proxyAddress, endpoint string) (*types.Settings, error) {
if settings == nil || settings.AccountID == "" {
return nil, status.Errorf(status.InvalidArgument, "account id is required")
}
if strings.TrimSpace(providerCluster) == "" {
return nil, fmt.Errorf("bootstrap settings: provider cluster is required")
if err := m.requirePermission(ctx, settings.AccountID, userID, modules.AgentNetworkSettings, operations.Create); err != nil {
return nil, err
}
existing, err := st.GetAgentNetworkSettings(ctx, store.LockingStrengthNone, accountID)
if err == nil {
return existing, nil
hasProxyAddress := strings.TrimSpace(proxyAddress) != ""
hasEndpoint := strings.TrimSpace(endpoint) != ""
if hasProxyAddress == hasEndpoint {
return nil, status.Errorf(status.InvalidArgument, "exactly one of proxy_address and endpoint is required")
}
if !isNotFound(err) {
// Fail fast on an existing row for a clean 409; the insert below stays
// the authority against concurrent bootstraps (the primary key wins).
if _, err := m.store.GetAgentNetworkSettings(ctx, store.LockingStrengthNone, settings.AccountID); err == nil {
return nil, status.Errorf(status.AlreadyExists, "agent network settings already bootstrapped for account %s", settings.AccountID)
} else if !isNotFound(err) {
return nil, fmt.Errorf("get agent network settings: %w", err)
}
siblings, err := st.GetAgentNetworkSettingsByCluster(ctx, store.LockingStrengthNone, providerCluster)
if err != nil {
return nil, fmt.Errorf("list agent network settings on cluster: %w", err)
}
taken := make(map[string]struct{}, len(siblings))
for _, s := range siblings {
taken[s.Subdomain] = struct{}{}
}
suffix := accountID
if len(suffix) > 4 {
suffix = suffix[:4]
}
m.labelRngMu.Lock()
subdomain := labelgen.PickUnique(m.labelRng, taken, suffix)
m.labelRngMu.Unlock()
now := time.Now().UTC()
settings := types.DefaultSettings(accountID)
settings.Cluster = providerCluster
settings.Subdomain = subdomain
settings.CreatedAt = now
settings.UpdatedAt = now
if err := st.SaveAgentNetworkSettings(ctx, settings); err != nil {
return nil, fmt.Errorf("save agent network settings: %w", err)
var err error
if hasEndpoint {
err = m.bootstrapSelfAddressed(ctx, settings, endpoint)
} else {
err = m.bootstrapLabeled(ctx, settings, proxyAddress)
}
if err != nil {
return nil, err
}
m.accountManager.StoreEvent(ctx, userID, settings.AccountID, settings.AccountID, activity.AgentNetworkSettingsUpdated, map[string]any{
"bootstrapped": true,
"endpoint": settings.Domain,
"dedicated": settings.Dedicated(),
})
m.reconcile(ctx, settings.AccountID)
return settings, nil
}
// bootstrapSelfAddressed claims the given hostname as the account's endpoint,
// served only by a proxy declaring exactly that address (Domain ==
// ProxyAddress). The domain unique index is the arbiter of availability.
func (m *managerImpl) bootstrapSelfAddressed(ctx context.Context, settings *types.Settings, endpoint string) error {
hostname, err := types.NormalizeHostname(endpoint)
if err != nil {
return status.Errorf(status.InvalidArgument, "invalid endpoint: %s", err)
}
settings.Domain = hostname
settings.ProxyAddress = hostname
if err := m.store.CreateAgentNetworkSettings(ctx, settings); err != nil {
if isUniqueConstraintError(err) {
// The violation is either the account primary key (a concurrent
// bootstrap for the same account won) or the domain index
// (another account holds the hostname). Distinguish by re-read.
if _, getErr := m.store.GetAgentNetworkSettings(ctx, store.LockingStrengthNone, settings.AccountID); getErr == nil {
return status.Errorf(status.AlreadyExists, "agent network settings already bootstrapped for account %s", settings.AccountID)
}
return status.Errorf(status.AlreadyExists, "endpoint %s is already taken", hostname)
}
return fmt.Errorf("create agent network settings: %w", err)
}
return nil
}
// bootstrapLabeled allocates a labeled endpoint one label beneath the given
// cluster address: Domain = <label>.<proxyAddress>, served by whichever proxy
// declares the parent. Labels are adjective-noun tuples; a candidate is
// checked by read and the domain unique index stays the authority, so a
// concurrent allocation of the same tuple surfaces as a unique violation and
// another tuple is drawn.
func (m *managerImpl) bootstrapLabeled(ctx context.Context, settings *types.Settings, proxyAddress string) error {
parent, err := types.NormalizeHostname(proxyAddress)
if err != nil {
return status.Errorf(status.InvalidArgument, "invalid proxy_address: %s", err)
}
for attempt := 1; attempt <= maxDomainAllocationAttempts; attempt++ {
m.labelRngMu.Lock()
label := labelgen.PickTuple(m.labelRng)
m.labelRngMu.Unlock()
if label == "" {
// Only reachable if either word pool were emptied. An empty label
// would produce a broken endpoint like ".example.com", so fail
// loudly rather than looping or inserting.
return fmt.Errorf("allocate agent network endpoint for account %s: label generator returned an empty label", settings.AccountID)
}
candidate, err := types.NormalizeHostname(label + "." + parent)
if err != nil {
return status.Errorf(status.InvalidArgument, "proxy_address leaves no room for a label: %s", err)
}
_, err = m.store.GetAgentNetworkSettingsByDomain(ctx, store.LockingStrengthNone, candidate)
if err == nil {
log.WithContext(ctx).Tracef("agent-network endpoint %q taken, retrying (attempt %d/%d)", candidate, attempt, maxDomainAllocationAttempts)
continue
}
if !isNotFound(err) {
return fmt.Errorf("check agent network endpoint availability: %w", err)
}
settings.Domain = candidate
settings.ProxyAddress = parent
if err := m.store.CreateAgentNetworkSettings(ctx, settings); err != nil {
if isUniqueConstraintError(err) {
// A concurrent bootstrap for the same account may have won on
// the primary key — return the conflict. A lost race on the
// domain index just means the tuple was taken between the
// read and the insert: draw another.
if _, getErr := m.store.GetAgentNetworkSettings(ctx, store.LockingStrengthNone, settings.AccountID); getErr == nil {
return status.Errorf(status.AlreadyExists, "agent network settings already bootstrapped for account %s", settings.AccountID)
}
log.WithContext(ctx).Tracef("agent-network endpoint %q lost an allocation race, retrying (attempt %d/%d)", candidate, attempt, maxDomainAllocationAttempts)
continue
}
return fmt.Errorf("create agent network settings: %w", err)
}
return nil
}
return fmt.Errorf("allocate agent network endpoint for account %s: %d attempts exhausted", settings.AccountID, maxDomainAllocationAttempts)
}
// isUniqueConstraintError reports whether err is a database unique-constraint
// violation, matched on the driver message because CreateAgentNetworkSettings
// deliberately returns the driver error unwrapped.
func isUniqueConstraintError(err error) bool {
if err == nil {
return false
}
msg := err.Error()
return strings.Contains(msg, "(SQLSTATE 23505)") || // postgres
strings.Contains(msg, "Error 1062 (23000)") || // mysql
strings.Contains(msg, "UNIQUE constraint failed") // sqlite
}
// ListConsumption returns every consumption row recorded for the
// account, ordered window-newest-first. Backs the dashboard's basic
// counter view; permission gate is the same Read role that gates
@@ -879,7 +1021,7 @@ func (*mockManager) GetProvider(_ context.Context, _, _, _ string) (*types.Provi
return &types.Provider{}, nil
}
func (*mockManager) CreateProvider(_ context.Context, _ string, p *types.Provider, _ string) (*types.Provider, error) {
func (*mockManager) CreateProvider(_ context.Context, _ string, p *types.Provider) (*types.Provider, error) {
return p, nil
}
@@ -947,10 +1089,23 @@ func (*mockManager) GetSettings(_ context.Context, accountID, _ string) (*types.
return types.DefaultSettings(accountID), nil
}
func (*mockManager) CreateSettings(_ context.Context, _ string, s *types.Settings, proxyAddress, endpoint string) (*types.Settings, error) {
if endpoint != "" {
s.Domain = endpoint
s.ProxyAddress = endpoint
} else {
s.Domain = "mock." + proxyAddress
s.ProxyAddress = proxyAddress
}
return s, nil
}
func (*mockManager) UpdateSettings(_ context.Context, _ string, s *types.Settings) (*types.Settings, error) {
return s, nil
}
func (*mockManager) DeleteSettings(_ context.Context, _, _ string) error { return nil }
func (*mockManager) ListConsumption(_ context.Context, _, _ string) ([]*types.Consumption, error) {
return nil, nil
}
@@ -1,134 +0,0 @@
package agentnetwork
import (
"context"
"runtime"
"testing"
"github.com/golang/mock/gomock"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
"github.com/netbirdio/netbird/management/internals/modules/agentnetwork/types"
"github.com/netbirdio/netbird/management/server/account"
"github.com/netbirdio/netbird/management/server/permissions"
"github.com/netbirdio/netbird/management/server/permissions/modules"
"github.com/netbirdio/netbird/management/server/permissions/operations"
"github.com/netbirdio/netbird/management/server/store"
nbtypes "github.com/netbirdio/netbird/management/server/types"
"github.com/netbirdio/netbird/shared/management/status"
)
// bootstrapFixture wires a real sqlite store to a gomock permissions manager
// so tests can grant the provider permission while denying (or never
// expecting) the settings one.
type bootstrapFixture struct {
manager Manager
store store.Store
perms *permissions.MockManager
}
func newBootstrapFixture(t *testing.T) *bootstrapFixture {
t.Helper()
if runtime.GOOS == "windows" {
t.Skip("sqlite store not properly supported on Windows yet")
}
t.Setenv("NETBIRD_STORE_ENGINE", string(nbtypes.SqliteStoreEngine))
st, cleanUp, err := store.NewTestStoreFromSQL(context.Background(), "", t.TempDir())
require.NoError(t, err, "test store setup must succeed")
t.Cleanup(cleanUp)
ctrl := gomock.NewController(t)
perms := permissions.NewMockManager(ctrl)
accounts := account.NewMockManager(ctrl)
accounts.EXPECT().StoreEvent(gomock.Any(), gomock.Any(), gomock.Any(), gomock.Any(), gomock.Any(), gomock.Any()).AnyTimes()
accounts.EXPECT().UpdateAccountPeers(gomock.Any(), gomock.Any(), gomock.Any()).AnyTimes()
accounts.EXPECT().BufferUpdateAccountPeers(gomock.Any(), gomock.Any(), gomock.Any()).AnyTimes()
return &bootstrapFixture{
manager: NewManager(st, perms, accounts, nil),
store: st,
perms: perms,
}
}
func (f *bootstrapFixture) expectPermission(accountID, userID string, module modules.Module, op operations.Operation, allowed bool) {
f.perms.EXPECT().
ValidateUserPermissions(gomock.Any(), accountID, userID, module, op).
Return(allowed, context.Background(), nil)
}
func newBootstrapProvider(accountID string) *types.Provider {
p := types.NewProvider(accountID)
p.Name = "openai"
p.UpstreamURL = "https://api.openai.com"
p.APIKey = "sk-test"
p.Enabled = true
return p
}
// TestCreateProviderBootstrapRequiresSettingsPermission pins the gate on the
// one-time settings bootstrap: creating the first provider with a
// bootstrap_cluster pins the account's cluster and subdomain, which is a
// settings write and must not ride on the providers permission alone.
func TestCreateProviderBootstrapRequiresSettingsPermission(t *testing.T) {
ctx := context.Background()
t.Run("denied without settings permission", func(t *testing.T) {
f := newBootstrapFixture(t)
f.expectPermission("account1", "user1", modules.AgentNetworkProviders, operations.Create, true)
f.expectPermission("account1", "user1", modules.AgentNetworkSettings, operations.Create, false)
_, err := f.manager.CreateProvider(ctx, "user1", newBootstrapProvider("account1"), "cluster1.example.com")
require.Error(t, err, "bootstrap without settings permission must fail")
var sErr *status.Error
require.ErrorAs(t, err, &sErr)
assert.Equal(t, status.PermissionDenied, sErr.Type(), "denial should surface as permission denied")
providers, err := f.store.GetAccountAgentNetworkProviders(ctx, store.LockingStrengthNone, "account1")
require.NoError(t, err)
assert.Empty(t, providers, "provider must not be persisted when bootstrap is denied")
_, err = f.store.GetAgentNetworkSettings(ctx, store.LockingStrengthNone, "account1")
assert.Error(t, err, "settings row must not be created when bootstrap is denied")
})
t.Run("allowed with settings permission", func(t *testing.T) {
f := newBootstrapFixture(t)
f.expectPermission("account1", "user1", modules.AgentNetworkProviders, operations.Create, true)
f.expectPermission("account1", "user1", modules.AgentNetworkSettings, operations.Create, true)
created, err := f.manager.CreateProvider(ctx, "user1", newBootstrapProvider("account1"), "cluster1.example.com")
require.NoError(t, err, "bootstrap with both permissions must succeed")
require.NotNil(t, created)
settings, err := f.store.GetAgentNetworkSettings(ctx, store.LockingStrengthNone, "account1")
require.NoError(t, err, "bootstrap must create the settings row")
assert.Equal(t, "cluster1.example.com", settings.Cluster, "settings should pin the bootstrap cluster")
})
t.Run("existing settings need no settings permission", func(t *testing.T) {
f := newBootstrapFixture(t)
require.NoError(t, f.store.SaveAgentNetworkSettings(ctx, &types.Settings{
AccountID: "account1",
Cluster: "cluster1.example.com",
Subdomain: "existing",
}), "pre-existing settings row setup must succeed")
// Only the providers permission may be consulted: gomock fails the
// test on any unexpected settings-permission call.
f.expectPermission("account1", "user1", modules.AgentNetworkProviders, operations.Create, true)
_, err := f.manager.CreateProvider(ctx, "user1", newBootstrapProvider("account1"), "cluster1.example.com")
require.NoError(t, err, "create with existing settings must not require the settings permission")
})
t.Run("no bootstrap cluster needs no settings permission", func(t *testing.T) {
f := newBootstrapFixture(t)
f.expectPermission("account1", "user1", modules.AgentNetworkProviders, operations.Create, true)
_, err := f.manager.CreateProvider(ctx, "user1", newBootstrapProvider("account1"), "")
require.NoError(t, err, "create without bootstrap must not require the settings permission")
})
}
@@ -10,6 +10,17 @@ import (
"github.com/netbirdio/netbird/shared/management/proto"
)
// syntheticMapping pairs a synthesised proxy mapping with the address of the
// proxy that serves it. The cluster is recorded rather than derived from the
// mapping's domain: ProxyMapping does not carry it, and the previous derivation
// -- everything after the first DNS label -- is wrong whenever the service's
// domain is not one label under its proxy's address, which silently addressed
// updates to a cluster no proxy declares.
type syntheticMapping struct {
mapping *proto.ProxyMapping
cluster string
}
// reconcile recomputes the synthesised reverse-proxy services for an
// account, diffs them against the previously-synthesised set in the
// in-memory cache, and emits Create / Update / Delete proxy mappings
@@ -45,18 +56,21 @@ func (m *managerImpl) reconcile(ctx context.Context, accountID string) {
}
oidcCfg := m.proxyController.GetOIDCValidationConfig()
current := make(map[string]*proto.ProxyMapping, len(services))
current := make(map[string]syntheticMapping, len(services))
for _, svc := range services {
if svc == nil || svc.ID == "" {
continue
}
current[svc.ID] = svc.ToProtoMapping(rpservice.Update, "", oidcCfg)
current[svc.ID] = syntheticMapping{
mapping: svc.ToProtoMapping(rpservice.Update, "", oidcCfg),
cluster: svc.ProxyCluster,
}
}
m.reconcileMu.Lock()
previous := m.reconcileCache[accountID]
if previous == nil {
previous = make(map[string]*proto.ProxyMapping)
previous = make(map[string]syntheticMapping)
}
creates, updates, deletes := diffMappings(previous, current)
@@ -67,34 +81,36 @@ func (m *managerImpl) reconcile(ctx context.Context, accountID string) {
}
m.reconcileMu.Unlock()
for _, mapping := range creates {
mapping.Type = proto.ProxyMappingUpdateType_UPDATE_TYPE_CREATED
m.proxyController.SendServiceUpdateToCluster(ctx, accountID, mapping, clusterFromMapping(mapping))
for _, entry := range creates {
entry.mapping.Type = proto.ProxyMappingUpdateType_UPDATE_TYPE_CREATED
m.proxyController.SendServiceUpdateToCluster(ctx, accountID, entry.mapping, entry.cluster)
}
for _, mapping := range updates {
mapping.Type = proto.ProxyMappingUpdateType_UPDATE_TYPE_MODIFIED
m.proxyController.SendServiceUpdateToCluster(ctx, accountID, mapping, clusterFromMapping(mapping))
for _, entry := range updates {
entry.mapping.Type = proto.ProxyMappingUpdateType_UPDATE_TYPE_MODIFIED
m.proxyController.SendServiceUpdateToCluster(ctx, accountID, entry.mapping, entry.cluster)
}
for _, mapping := range deletes {
mapping.Type = proto.ProxyMappingUpdateType_UPDATE_TYPE_REMOVED
m.proxyController.SendServiceUpdateToCluster(ctx, accountID, mapping, clusterFromMapping(mapping))
for _, entry := range deletes {
entry.mapping.Type = proto.ProxyMappingUpdateType_UPDATE_TYPE_REMOVED
m.proxyController.SendServiceUpdateToCluster(ctx, accountID, entry.mapping, entry.cluster)
}
}
// diffMappings classifies the previous→current transition for a
// single account into Create / Update / Delete sets.
// diffMappings classifies the previous→current transition for a single
// account into Create / Update / Delete sets.
//
// Cluster moves (current.cluster != previous.cluster) are surfaced as
// a Delete on the old cluster + Create on the new — handled by
// emitting both a delete (on previous mapping) and a create (on the
// current mapping) for that service ID.
func diffMappings(previous, current map[string]*proto.ProxyMapping) (creates, updates, deletes []*proto.ProxyMapping) {
// A change of serving proxy for the same service ID is surfaced as a Delete
// addressed to the old proxy plus a Create addressed to the new one, so the
// mapping actually moves. Comparing the recorded cluster is what makes that
// detectable: with a placement-free endpoint the mapping's domain is identical
// before and after the move, so nothing about the mapping itself reveals it.
func diffMappings(previous, current map[string]syntheticMapping) (creates, updates, deletes []syntheticMapping) {
for id, cur := range current {
prev, existed := previous[id]
switch {
case !existed:
creates = append(creates, cur)
case prev.GetDomain() == "" || cur.GetAccountId() == prev.GetAccountId() && currentClusterChanged(prev, cur):
case prev.mapping.GetDomain() == "" ||
cur.mapping.GetAccountId() == prev.mapping.GetAccountId() && prev.cluster != cur.cluster:
deletes = append(deletes, prev)
creates = append(creates, cur)
default:
@@ -108,24 +124,3 @@ func diffMappings(previous, current map[string]*proto.ProxyMapping) (creates, up
}
return creates, updates, deletes
}
func currentClusterChanged(prev, cur *proto.ProxyMapping) bool {
return clusterFromMapping(prev) != clusterFromMapping(cur)
}
// clusterFromMapping returns the cluster the mapping should be sent
// to. ProxyMapping doesn't carry the cluster directly, so we rely on
// the synthesised service's domain (`<slug>.<cluster>`) and split on
// the first '.'.
func clusterFromMapping(m *proto.ProxyMapping) string {
if m == nil {
return ""
}
domain := m.GetDomain()
for i := 0; i < len(domain); i++ {
if domain[i] == '.' {
return domain[i+1:]
}
}
return ""
}
@@ -21,7 +21,7 @@ func newReconcileMgr(t *testing.T, ctrl *gomock.Controller) (*managerImpl, *stor
return &managerImpl{
store: mockStore,
proxyController: mockProxy,
reconcileCache: make(map[string]map[string]*proto.ProxyMapping),
reconcileCache: make(map[string]map[string]syntheticMapping),
}, mockStore, mockProxy
}
@@ -52,9 +52,9 @@ func newReconcileTestPolicy(providerID, sourceGroupID string) *types.Policy {
func newReconcileTestSettings() *types.Settings {
return &types.Settings{
AccountID: "acct-1",
Cluster: "eu.proxy.netbird.io",
Subdomain: "violet",
AccountID: "acct-1",
Domain: "violet.eu.proxy.netbird.io",
ProxyAddress: "eu.proxy.netbird.io",
}
}
@@ -196,7 +196,7 @@ func TestReconcile_PolicyRemoved_EmitsDelete(t *testing.T) {
func TestReconcile_NilProxyController_NoOp(t *testing.T) {
ctx := context.Background()
mgr := &managerImpl{
reconcileCache: make(map[string]map[string]*proto.ProxyMapping),
reconcileCache: make(map[string]map[string]syntheticMapping),
}
// Must not panic; must not query the store.
mgr.reconcile(ctx, "acct-1")
@@ -212,21 +212,78 @@ func TestReconcile_EmptyAccountID_NoOp(t *testing.T) {
mgr.reconcile(ctx, "")
}
func TestClusterFromMapping(t *testing.T) {
tests := []struct {
name string
domain string
want string
}{
{"simple", "openai.eu.proxy.netbird.io", "eu.proxy.netbird.io"},
{"deeply nested", "a.b.c.d", "b.c.d"},
{"no dot", "openai", ""},
{"empty", "", ""},
// TestDiffMappings_ServingProxyChange — when the proxy serving an account
// changes, the same service ID must be deleted on the old proxy and created on
// the new one. The cluster cannot be recovered from the mapping's domain: with a
// placement-free endpoint the domain does not change at all when the serving
// proxy does, so a domain-derived cluster sees no change and emits a plain
// update, addressed to a proxy that does not exist.
func TestDiffMappings_ServingProxyChange(t *testing.T) {
previous := map[string]syntheticMapping{
"svc-1": {
mapping: &proto.ProxyMapping{Id: "svc-1", AccountId: "acct-1", Domain: "brave-otter.gateway.example.com"},
cluster: "proxy.example.com",
},
}
for _, tt := range tests {
t.Run(tt.name, func(t *testing.T) {
got := clusterFromMapping(&proto.ProxyMapping{Domain: tt.domain})
assert.Equal(t, tt.want, got)
})
current := map[string]syntheticMapping{
"svc-1": {
mapping: &proto.ProxyMapping{Id: "svc-1", AccountId: "acct-1", Domain: "brave-otter.gateway.example.com"},
cluster: "brave-otter.gateway.example.com",
},
}
creates, updates, deletes := diffMappings(previous, current)
if assert.Len(t, deletes, 1, "the old proxy must be told to drop the mapping") {
assert.Equal(t, "proxy.example.com", deletes[0].cluster)
}
if assert.Len(t, creates, 1, "the new proxy must be told to add it") {
assert.Equal(t, "brave-otter.gateway.example.com", creates[0].cluster)
}
assert.Empty(t, updates, "a serving-proxy move is a delete plus a create, not an update")
}
// TestDiffMappings_UnchangedClusterIsAnUpdate keeps the ordinary path: same
// service, same proxy, changed contents.
func TestDiffMappings_UnchangedClusterIsAnUpdate(t *testing.T) {
previous := map[string]syntheticMapping{
"svc-1": {
mapping: &proto.ProxyMapping{Id: "svc-1", AccountId: "acct-1", Domain: "otter.proxy.example.com"},
cluster: "proxy.example.com",
},
}
current := map[string]syntheticMapping{
"svc-1": {
mapping: &proto.ProxyMapping{Id: "svc-1", AccountId: "acct-1", Domain: "otter.proxy.example.com"},
cluster: "proxy.example.com",
},
}
creates, updates, deletes := diffMappings(previous, current)
assert.Empty(t, creates)
assert.Empty(t, deletes)
if assert.Len(t, updates, 1) {
assert.Equal(t, "proxy.example.com", updates[0].cluster)
}
}
// TestDiffMappings_RemovedServiceIsDeletedOnItsOwnCluster — a service that has
// gone away is deleted on the cluster it was last served by, which is recorded
// rather than re-derived.
func TestDiffMappings_RemovedServiceIsDeletedOnItsOwnCluster(t *testing.T) {
previous := map[string]syntheticMapping{
"svc-1": {
mapping: &proto.ProxyMapping{Id: "svc-1", AccountId: "acct-1", Domain: "brave-otter.gateway.example.com"},
cluster: "brave-otter.gateway.example.com",
},
}
creates, updates, deletes := diffMappings(previous, map[string]syntheticMapping{})
assert.Empty(t, creates)
assert.Empty(t, updates)
if assert.Len(t, deletes, 1) {
assert.Equal(t, "brave-otter.gateway.example.com", deletes[0].cluster)
}
}
@@ -0,0 +1,225 @@
package agentnetwork
import (
"context"
"runtime"
"strings"
"testing"
"github.com/golang/mock/gomock"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
"github.com/netbirdio/netbird/management/internals/modules/agentnetwork/types"
"github.com/netbirdio/netbird/management/server/account"
"github.com/netbirdio/netbird/management/server/permissions"
"github.com/netbirdio/netbird/management/server/permissions/modules"
"github.com/netbirdio/netbird/management/server/permissions/operations"
"github.com/netbirdio/netbird/management/server/store"
nbtypes "github.com/netbirdio/netbird/management/server/types"
"github.com/netbirdio/netbird/shared/management/status"
)
// bootstrapFixture wires a real sqlite store to a gomock permissions manager
// so tests can grant or deny the settings permission per case.
type bootstrapFixture struct {
manager Manager
store store.Store
perms *permissions.MockManager
}
func newBootstrapFixture(t *testing.T) *bootstrapFixture {
t.Helper()
if runtime.GOOS == "windows" {
t.Skip("sqlite store not properly supported on Windows yet")
}
t.Setenv("NETBIRD_STORE_ENGINE", string(nbtypes.SqliteStoreEngine))
st, cleanUp, err := store.NewTestStoreFromSQL(context.Background(), "", t.TempDir())
require.NoError(t, err, "test store setup must succeed")
t.Cleanup(cleanUp)
ctrl := gomock.NewController(t)
perms := permissions.NewMockManager(ctrl)
accounts := account.NewMockManager(ctrl)
accounts.EXPECT().StoreEvent(gomock.Any(), gomock.Any(), gomock.Any(), gomock.Any(), gomock.Any(), gomock.Any()).AnyTimes()
accounts.EXPECT().UpdateAccountPeers(gomock.Any(), gomock.Any(), gomock.Any()).AnyTimes()
accounts.EXPECT().BufferUpdateAccountPeers(gomock.Any(), gomock.Any(), gomock.Any()).AnyTimes()
return &bootstrapFixture{
manager: NewManager(st, perms, accounts, nil),
store: st,
perms: perms,
}
}
func (f *bootstrapFixture) expectPermission(accountID, userID string, module modules.Module, op operations.Operation, allowed bool) {
f.perms.EXPECT().
ValidateUserPermissions(gomock.Any(), accountID, userID, module, op).
Return(allowed, context.Background(), nil)
}
func (f *bootstrapFixture) createSettings(ctx context.Context, accountID, userID, proxyAddress, endpoint string) (*types.Settings, error) {
return f.manager.CreateSettings(ctx, userID, types.DefaultSettings(accountID), proxyAddress, endpoint)
}
// TestCreateSettingsRequiresPermission pins the gate: bootstrap assigns the
// account's immutable endpoint, a settings write requiring the settings
// Create permission — and a denial leaves no row behind.
func TestCreateSettingsRequiresPermission(t *testing.T) {
ctx := context.Background()
f := newBootstrapFixture(t)
f.expectPermission("account1", "user1", modules.AgentNetworkSettings, operations.Create, false)
_, err := f.createSettings(ctx, "account1", "user1", "cluster1.example.com", "")
require.Error(t, err, "bootstrap without the settings permission must fail")
var sErr *status.Error
require.ErrorAs(t, err, &sErr)
assert.Equal(t, status.PermissionDenied, sErr.Type(), "denial should surface as permission denied")
_, err = f.store.GetAgentNetworkSettings(ctx, store.LockingStrengthNone, "account1")
assert.Error(t, err, "settings row must not be created when bootstrap is denied")
}
// TestCreateSettingsLabeled pins the labeled shape: the server allocates an
// adjective-noun label beneath the proxy address, the pin is not dedicated,
// and the domain records the full endpoint hostname.
func TestCreateSettingsLabeled(t *testing.T) {
ctx := context.Background()
f := newBootstrapFixture(t)
f.expectPermission("account1", "user1", modules.AgentNetworkSettings, operations.Create, true)
created, err := f.createSettings(ctx, "account1", "user1", "Cluster1.Example.com", "")
require.NoError(t, err, "labeled bootstrap must succeed")
assert.Equal(t, "cluster1.example.com", created.ProxyAddress, "proxy address must be pinned lowercased")
require.True(t, strings.HasSuffix(created.Domain, ".cluster1.example.com"),
"domain must hang one label beneath the proxy address: %s", created.Domain)
label := strings.TrimSuffix(created.Domain, ".cluster1.example.com")
assert.NotContains(t, label, ".", "the allocated label must be a single DNS label: %s", label)
assert.False(t, created.Dedicated(), "a labeled pin is not dedicated")
assert.Equal(t, created.Domain, created.Endpoint(), "the endpoint is the domain column")
stored, err := f.store.GetAgentNetworkSettings(ctx, store.LockingStrengthNone, "account1")
require.NoError(t, err, "bootstrap must persist the row")
assert.Equal(t, created.Domain, stored.Domain)
assert.Equal(t, created.ProxyAddress, stored.ProxyAddress)
}
// TestCreateSettingsSelfAddressed pins the dedicated shape: the endpoint is
// claimed verbatim (normalized), Domain == ProxyAddress, and the claim
// succeeds with no proxy declaring the address yet (address-first).
func TestCreateSettingsSelfAddressed(t *testing.T) {
ctx := context.Background()
f := newBootstrapFixture(t)
f.expectPermission("account1", "user1", modules.AgentNetworkSettings, operations.Create, true)
created, err := f.createSettings(ctx, "account1", "user1", "", "Brave-Otter.GW.Example.com")
require.NoError(t, err, "self-addressed bootstrap must succeed")
assert.Equal(t, "brave-otter.gw.example.com", created.Domain, "endpoint must be claimed lowercased")
assert.Equal(t, created.Domain, created.ProxyAddress, "self-addressed: proxy address is the endpoint")
assert.True(t, created.Dedicated(), "a self-addressed pin is dedicated")
}
// TestCreateSettingsIdentityFieldValidation pins the request contract: exactly
// one of proxyAddress and endpoint, and both must be well-formed hostnames.
func TestCreateSettingsIdentityFieldValidation(t *testing.T) {
ctx := context.Background()
cases := map[string]struct {
proxyAddress string
endpoint string
}{
"neither": {"", ""},
"both": {"cluster1.example.com", "gw.example.com"},
"trailing dot endpoint": {"", "gw.example.com."},
"leading dot endpoint": {"", ".gw.example.com"},
"whitespace inside": {"", "g w.example.com"},
"empty label in parent": {"eu..example.com", ""},
"hyphen-edged label": {"", "-gw.example.com"},
}
for name, tc := range cases {
t.Run(name, func(t *testing.T) {
f := newBootstrapFixture(t)
f.expectPermission("account1", "user1", modules.AgentNetworkSettings, operations.Create, true)
_, err := f.createSettings(ctx, "account1", "user1", tc.proxyAddress, tc.endpoint)
require.Error(t, err, "invalid identity input must be rejected")
var sErr *status.Error
require.ErrorAs(t, err, &sErr)
assert.Equal(t, status.InvalidArgument, sErr.Type(), "rejection must be a validation error")
_, err = f.store.GetAgentNetworkSettings(ctx, store.LockingStrengthNone, "account1")
assert.Error(t, err, "no row may be left behind by a rejected bootstrap")
})
}
}
// TestCreateSettingsConflictsOnSecondBootstrap pins that bootstrap is a
// one-time create per account: a second call is a conflict, whatever shape it
// asks for, and the original row survives untouched.
func TestCreateSettingsConflictsOnSecondBootstrap(t *testing.T) {
ctx := context.Background()
f := newBootstrapFixture(t)
f.expectPermission("account1", "user1", modules.AgentNetworkSettings, operations.Create, true)
first, err := f.createSettings(ctx, "account1", "user1", "cluster1.example.com", "")
require.NoError(t, err)
f.expectPermission("account1", "user1", modules.AgentNetworkSettings, operations.Create, true)
_, err = f.createSettings(ctx, "account1", "user1", "", "other.example.com")
require.Error(t, err, "second bootstrap must fail")
var sErr *status.Error
require.ErrorAs(t, err, &sErr)
assert.Equal(t, status.AlreadyExists, sErr.Type(), "second bootstrap must surface as a conflict")
stored, err := f.store.GetAgentNetworkSettings(ctx, store.LockingStrengthNone, "account1")
require.NoError(t, err)
assert.Equal(t, first.Domain, stored.Domain, "the original endpoint must survive the rejected bootstrap")
}
// TestCreateSettingsEndpointTaken pins global hostname uniqueness: a hostname
// held by one account cannot be claimed by another, in either direction —
// self-addressed onto self-addressed, or self-addressed onto an allocated
// labeled endpoint.
func TestCreateSettingsEndpointTaken(t *testing.T) {
ctx := context.Background()
f := newBootstrapFixture(t)
f.expectPermission("account1", "user1", modules.AgentNetworkSettings, operations.Create, true)
first, err := f.createSettings(ctx, "account1", "user1", "", "gw.example.com")
require.NoError(t, err)
f.expectPermission("account2", "user2", modules.AgentNetworkSettings, operations.Create, true)
_, err = f.createSettings(ctx, "account2", "user2", "", "gw.example.com")
require.Error(t, err, "a taken hostname must be refused")
var sErr *status.Error
require.ErrorAs(t, err, &sErr)
assert.Equal(t, status.AlreadyExists, sErr.Type(), "the refusal must surface as a conflict")
f.expectPermission("account3", "user3", modules.AgentNetworkSettings, operations.Create, true)
_, err = f.createSettings(ctx, "account3", "user3", "", first.Domain)
require.Error(t, err, "claiming another account's endpoint must be refused")
}
// TestCreateProviderHasNoSettingsSideEffects pins the decoupling: provider
// create needs only the providers permission (gomock fails the test on any
// settings-permission call) and never creates a settings row.
func TestCreateProviderHasNoSettingsSideEffects(t *testing.T) {
ctx := context.Background()
f := newBootstrapFixture(t)
f.expectPermission("account1", "user1", modules.AgentNetworkProviders, operations.Create, true)
provider := types.NewProvider("account1")
provider.Name = "openai"
provider.UpstreamURL = "https://api.openai.com"
provider.APIKey = "sk-test"
provider.Enabled = true
created, err := f.manager.CreateProvider(ctx, "user1", provider)
require.NoError(t, err, "provider create must succeed on the providers permission alone")
require.NotNil(t, created)
_, err = f.store.GetAgentNetworkSettings(ctx, store.LockingStrengthNone, "account1")
assert.Error(t, err, "provider create must not conjure a settings row")
}
@@ -89,7 +89,7 @@ func SynthesizeServicesForCluster(ctx context.Context, s store.Store, clusterAdd
return nil, nil
}
settingsRows, err := s.GetAgentNetworkSettingsByCluster(ctx, store.LockingStrengthNone, clusterAddr)
settingsRows, err := s.GetAgentNetworkSettingsByProxyAddress(ctx, store.LockingStrengthNone, clusterAddr)
if err != nil {
return nil, fmt.Errorf("list agent network settings on cluster: %w", err)
}
@@ -116,53 +116,41 @@ func SynthesizeServicesForCluster(ctx context.Context, s store.Store, clusterAdd
}
// SynthesizeServiceForDomain resolves a single agent-network service by its
// public endpoint domain. It lists the (few) settings rows on the domain's
// cluster, matches the one whose endpoint equals the domain, and synthesises
// only that account — avoiding full per-account synthesis for every tenant on
// the cluster, which is what auth/session paths previously paid. Returns nil
// (no error) when no account owns the domain.
// public endpoint domain — a point query on the settings domain unique index,
// then synthesis of just that account. Returns nil (no error) when no account
// owns the domain.
func SynthesizeServiceForDomain(ctx context.Context, s store.Store, domain string) (*rpservice.Service, error) {
domain = strings.TrimSpace(domain)
cluster := clusterFromDomain(domain)
if domain != "" && cluster != "" {
settingsRows, err := s.GetAgentNetworkSettingsByCluster(ctx, store.LockingStrengthNone, cluster)
if err != nil {
return nil, fmt.Errorf("list agent network settings on cluster: %w", err)
domain = strings.ToLower(strings.TrimSpace(domain))
if domain == "" {
return nil, nil //nolint:nilnil // optional lookup: no account owns the domain
}
settings, err := s.GetAgentNetworkSettingsByDomain(ctx, store.LockingStrengthNone, domain)
if err != nil {
if isNotFound(err) {
return nil, nil //nolint:nilnil // optional lookup: no account owns the domain
}
for _, settings := range settingsRows {
if settings == nil || settings.Endpoint() != domain {
continue
}
services, serr := SynthesizeServices(ctx, s, settings.AccountID)
if serr != nil {
return nil, serr
}
for _, svc := range services {
if svc != nil && svc.Domain == domain {
return svc, nil
}
}
break
return nil, fmt.Errorf("get agent network settings by domain: %w", err)
}
services, err := SynthesizeServices(ctx, s, settings.AccountID)
if err != nil {
return nil, err
}
for _, svc := range services {
if svc != nil && svc.Domain == domain {
return svc, nil
}
}
return nil, nil //nolint:nilnil // optional lookup: no account owns the domain
}
// clusterFromDomain returns the cluster portion of an endpoint domain (every
// label after the first).
func clusterFromDomain(domain string) string {
if i := strings.IndexByte(domain, '.'); i >= 0 {
return domain[i+1:]
}
return ""
}
// SynthesizeServices builds the in-memory reverse-proxy service that
// fronts the account's agent-network gateway. Returns nil when the
// account has no settings row, no enabled providers, or no enabled
// policies — in any of those cases there's nothing useful to expose.
//
// One service per (account, settings.Cluster) is emitted. The router
// One service per (account, settings.ProxyAddress) is emitted. The router
// middleware encodes a denormalised model→provider routing table
// (auth headers + decrypted API keys baked in); the policy_check
// middleware encodes per-provider authorised group IDs derived from
@@ -175,7 +163,7 @@ func SynthesizeServices(ctx context.Context, s store.Store, accountID string) ([
if err != nil {
return nil, err
}
if !ok || strings.TrimSpace(settings.Cluster) == "" {
if !ok || strings.TrimSpace(settings.ProxyAddress) == "" {
return nil, nil
}
@@ -934,7 +922,7 @@ func buildAccountService(
middlewares []rpservice.MiddlewareConfig,
sessionPriv, sessionPub string,
) *rpservice.Service {
cluster := settings.Cluster
cluster := settings.ProxyAddress
domain := settings.Endpoint()
serviceID := SynthesizedServiceIDPrefix + accountID
@@ -147,7 +147,7 @@ func TestReconcile_RealStore_PushesPrivateAfterStatusToggle(t *testing.T) {
store: s,
accountManager: noopAccountManager{},
proxyController: ctrl,
reconcileCache: make(map[string]map[string]*proto.ProxyMapping),
reconcileCache: make(map[string]map[string]syntheticMapping),
}
m.reconcile(ctx, testAccountID) // initial, provider enabled
@@ -19,15 +19,14 @@ import (
const (
testAccountID = "acct-1"
testCluster = "eu.proxy.netbird.io"
testSubdomain = "violet"
testEndpoint = "violet.eu.proxy.netbird.io"
)
func newSynthTestSettings() *types.Settings {
return &types.Settings{
AccountID: testAccountID,
Cluster: testCluster,
Subdomain: testSubdomain,
AccountID: testAccountID,
Domain: testEndpoint,
ProxyAddress: testCluster,
}
}
@@ -1,6 +1,7 @@
package types
import (
"fmt"
"strings"
"time"
@@ -12,13 +13,23 @@ import (
// the long-term aggregate and are retained independently.
const DefaultAccessLogRetentionDays = 30
// Settings is the per-account agent-network configuration row. One
// row per account. Cluster + Subdomain are immutable once written and
// produce the public endpoint agents call (`<subdomain>.<cluster>`).
// Settings is the per-account agent-network configuration row. One row per
// account. Domain and ProxyAddress are assigned at bootstrap and immutable
// thereafter; a persisted row is always fully allocated — there is no "row
// exists, endpoint pending" state.
type Settings struct {
AccountID string `gorm:"primaryKey"`
Cluster string
Subdomain string `gorm:"index:idx_agent_network_settings_cluster_subdomain"`
// Domain is the gateway endpoint hostname agents call. Globally unique
// across accounts. Sized explicitly because MySQL cannot index an
// unbounded TEXT column; 255 covers the RFC 1035 253-octet bound.
Domain string `gorm:"type:varchar(255);uniqueIndex:idx_agent_network_settings_domain"`
// ProxyAddress is the declared cluster address of the proxy serving this
// account's gateway. Either equal to Domain — a proxy dedicated to this
// account, declaring the tenant's own hostname — or Domain's immediate
// parent, with the endpoint one label beneath it on a shared cluster.
ProxyAddress string `gorm:"type:varchar(255);index:idx_agent_network_settings_proxy_address"`
// Account-level collection controls sourced by the synthesizer.
// EnableLogCollection gates the per-request access-log trail and defaults
@@ -45,9 +56,9 @@ func (Settings) TableName() string { return "agent_network_settings" }
// DefaultSettings returns the settings an account observes before its row is
// bootstrapped: log collection on with the default retention, everything else
// off, and no cluster/subdomain assigned yet. Bootstrap persists exactly these
// values plus the assigned cluster and subdomain, so the pre-bootstrap read
// and the freshly bootstrapped row agree.
// off, and no domain or proxy address assigned yet. Bootstrap persists exactly
// these values plus the assigned domain and proxy address, so the
// pre-bootstrap read and the freshly bootstrapped row agree.
func DefaultSettings(accountID string) *Settings {
return &Settings{
AccountID: accountID,
@@ -56,14 +67,15 @@ func DefaultSettings(accountID string) *Settings {
}
}
// Endpoint returns the bare hostname agents reach this account at:
// `<subdomain>.<cluster>`. Empty until both halves are assigned at bootstrap.
func (s *Settings) Endpoint() string {
if s.Cluster == "" || s.Subdomain == "" {
return ""
}
return s.Subdomain + "." + s.Cluster
}
// Endpoint returns the bare hostname agents reach this account at — the
// Domain column. Empty until the row is bootstrapped.
func (s *Settings) Endpoint() string { return s.Domain }
// Dedicated reports whether the account's gateway is served by a proxy
// dedicated to it — the self-addressed shape, where the serving proxy declares
// the endpoint hostname itself. The alternative (labeled) shape has the
// endpoint one label beneath a shared cluster's address.
func (s *Settings) Dedicated() bool { return s.Domain != "" && s.Domain == s.ProxyAddress }
// ToAPIResponse renders the settings as the API representation. The
// timestamps are omitted while zero — a default (not yet bootstrapped) view
@@ -71,9 +83,9 @@ func (s *Settings) Endpoint() string {
func (s *Settings) ToAPIResponse() *api.AgentNetworkSettings {
retention := s.AccessLogRetentionDays
resp := &api.AgentNetworkSettings{
Cluster: s.Cluster,
Subdomain: s.Subdomain,
Endpoint: s.Endpoint(),
ProxyAddress: s.ProxyAddress,
Dedicated: s.Dedicated(),
EnableLogCollection: s.EnableLogCollection,
EnablePromptCollection: s.EnablePromptCollection,
RedactPii: s.RedactPii,
@@ -90,19 +102,91 @@ func (s *Settings) ToAPIResponse() *api.AgentNetworkSettings {
return resp
}
// FromAPIRequest applies the request onto the receiver. The mutable
// collection fields are always replaced with the request values. Cluster
// participates only in bootstrap and the immutability check (see
// Manager.UpdateSettings); Subdomain is server-assigned and never taken
// from a request.
// FromAPIRequest applies the update request onto the receiver: every mutable
// field is replaced with the request value, and the identity fields (Domain,
// ProxyAddress) carry the request's echo of the assigned values. The identity
// fields are never written to the stored row — UpdateSettings compares them
// against it and rejects the request when they differ, so PUT keeps the
// house convention of requiring every field while the endpoint and proxy
// address stay immutable.
//
// Every field is required by the schema, so none is presence-sensitive.
// AccessLogRetentionDays in particular must stay required: the caller receives
// a zero-valued Settings, and UpdateSettings copies each field onto the stored
// row unconditionally, so an omitted value would be written as 0 — which the
// API documents as "keep indefinitely". Making retention optional would
// therefore let a client silently maximise log retention by leaving it out.
func (s *Settings) FromAPIRequest(req *api.AgentNetworkSettingsRequest) {
if req.Cluster != nil {
s.Cluster = strings.TrimSpace(*req.Cluster)
}
s.Domain = req.Endpoint
s.ProxyAddress = req.ProxyAddress
s.EnableLogCollection = req.EnableLogCollection
s.EnablePromptCollection = req.EnablePromptCollection
s.RedactPii = req.RedactPii
s.AccessLogRetentionDays = req.AccessLogRetentionDays
}
// FromAPICreateRequest applies the optional collection toggles of a bootstrap
// request onto the receiver (typically DefaultSettings), leaving defaults in
// place for omitted fields. The identity fields are resolved by the manager
// from the request's proxy_address / endpoint, not copied here.
func (s *Settings) FromAPICreateRequest(req *api.AgentNetworkSettingsCreateRequest) {
if req.EnableLogCollection != nil {
s.EnableLogCollection = *req.EnableLogCollection
}
if req.EnablePromptCollection != nil {
s.EnablePromptCollection = *req.EnablePromptCollection
}
if req.RedactPii != nil {
s.RedactPii = *req.RedactPii
}
if req.AccessLogRetentionDays != nil {
s.AccessLogRetentionDays = *req.AccessLogRetentionDays
}
}
// maxHostnameLength is the RFC 1035 bound on a full domain name.
const maxHostnameLength = 253
// NormalizeHostname lowercases and trims a caller-supplied hostname and
// validates its shape: non-empty DNS labels of letters, digits and inner
// hyphens, joined by single dots, within length bounds. Shapes that
// canonicalization cannot repair — leading/trailing dots, empty labels,
// whitespace inside the name — are rejected rather than guessed at, because
// the value lands in an immutable column.
func NormalizeHostname(raw string) (string, error) {
hostname := strings.ToLower(strings.TrimSpace(raw))
if hostname == "" {
return "", fmt.Errorf("hostname is empty")
}
if len(hostname) > maxHostnameLength {
return "", fmt.Errorf("hostname exceeds %d characters", maxHostnameLength)
}
for _, label := range strings.Split(hostname, ".") {
if err := validateHostnameLabel(label); err != nil {
return "", fmt.Errorf("invalid hostname %q: %w", hostname, err)
}
}
return hostname, nil
}
func validateHostnameLabel(label string) error {
if label == "" {
return fmt.Errorf("empty label (leading, trailing or doubled dot)")
}
if len(label) > 63 {
return fmt.Errorf("label %q exceeds 63 characters", label)
}
if label[0] == '-' || label[len(label)-1] == '-' {
return fmt.Errorf("label %q must not start or end with a hyphen", label)
}
for _, r := range label {
switch {
case r >= 'a' && r <= 'z':
case r >= '0' && r <= '9':
case r == '-':
default:
return fmt.Errorf("label %q contains invalid character %q", label, r)
}
}
return nil
}