mirror of
https://github.com/netbirdio/netbird.git
synced 2026-09-18 12:49:07 +02:00
[management, proxy] Add prompt-cache token and cost accounting to agent network usage (#6900)
Co-authored-by: braginini <bangvalo@gmail.com>
This commit is contained in:
Maycon Santos
co-authored by
braginini
braginini
parent
1e5b0a5c89
commit
4f6247b5c3
@@ -5822,13 +5822,48 @@ components:
|
||||
total_tokens:
|
||||
type: integer
|
||||
format: int64
|
||||
description: Total tokens consumed.
|
||||
description: Total tokens consumed, including prompt-cache tokens.
|
||||
example: 1840
|
||||
cached_input_tokens:
|
||||
type: integer
|
||||
format: int64
|
||||
description: Input tokens read from the provider's prompt cache. Additive to input_tokens for Anthropic-shape providers; a subset of input_tokens for OpenAI.
|
||||
example: 0
|
||||
cache_creation_tokens:
|
||||
type: integer
|
||||
format: int64
|
||||
description: Input tokens written to the provider's prompt cache. Zero for providers without a cache-write bucket.
|
||||
example: 30528
|
||||
cost_usd:
|
||||
type: number
|
||||
format: double
|
||||
description: Estimated USD cost of the request.
|
||||
example: 0.0231
|
||||
input_cost_usd:
|
||||
type: number
|
||||
format: double
|
||||
description: Cost of the non-cached input tokens. Base component of cost_usd.
|
||||
example: 0.0048
|
||||
cached_input_cost_usd:
|
||||
type: number
|
||||
format: double
|
||||
description: Cost of the prompt-cache read tokens. Base component of cost_usd, and part of cache_cost_usd.
|
||||
example: 0.0015
|
||||
cache_creation_cost_usd:
|
||||
type: number
|
||||
format: double
|
||||
description: Cost of the prompt-cache write tokens. Base component of cost_usd, and part of cache_cost_usd.
|
||||
example: 0.1130
|
||||
output_cost_usd:
|
||||
type: number
|
||||
format: double
|
||||
description: Cost of the output tokens. Base component of cost_usd.
|
||||
example: 0.0038
|
||||
cache_cost_usd:
|
||||
type: number
|
||||
format: double
|
||||
description: Portion of cost_usd billed for prompt-cache usage.
|
||||
example: 0.1145
|
||||
stream:
|
||||
type: boolean
|
||||
description: Whether the request was a streaming completion.
|
||||
@@ -5852,7 +5887,14 @@ components:
|
||||
- input_tokens
|
||||
- output_tokens
|
||||
- total_tokens
|
||||
- cached_input_tokens
|
||||
- cache_creation_tokens
|
||||
- input_cost_usd
|
||||
- cached_input_cost_usd
|
||||
- cache_creation_cost_usd
|
||||
- output_cost_usd
|
||||
- cost_usd
|
||||
- cache_cost_usd
|
||||
AgentNetworkAccessLogsResponse:
|
||||
type: object
|
||||
properties:
|
||||
@@ -5926,13 +5968,48 @@ components:
|
||||
total_tokens:
|
||||
type: integer
|
||||
format: int64
|
||||
description: Total tokens across the session.
|
||||
description: Total tokens across the session, including prompt-cache tokens.
|
||||
example: 12880
|
||||
cached_input_tokens:
|
||||
type: integer
|
||||
format: int64
|
||||
description: Total prompt-cache read tokens across the session.
|
||||
example: 0
|
||||
cache_creation_tokens:
|
||||
type: integer
|
||||
format: int64
|
||||
description: Total prompt-cache write tokens across the session.
|
||||
example: 30528
|
||||
cost_usd:
|
||||
type: number
|
||||
format: double
|
||||
description: Total estimated USD cost across the session.
|
||||
example: 0.1617
|
||||
input_cost_usd:
|
||||
type: number
|
||||
format: double
|
||||
description: Total cost of non-cached input tokens across the session.
|
||||
example: 0.0210
|
||||
cached_input_cost_usd:
|
||||
type: number
|
||||
format: double
|
||||
description: Total cost of prompt-cache read tokens across the session.
|
||||
example: 0.0015
|
||||
cache_creation_cost_usd:
|
||||
type: number
|
||||
format: double
|
||||
description: Total cost of prompt-cache write tokens across the session.
|
||||
example: 0.1130
|
||||
output_cost_usd:
|
||||
type: number
|
||||
format: double
|
||||
description: Total cost of output tokens across the session.
|
||||
example: 0.0262
|
||||
cache_cost_usd:
|
||||
type: number
|
||||
format: double
|
||||
description: Portion of cost_usd billed for prompt-cache usage across the session.
|
||||
example: 0.1145
|
||||
providers:
|
||||
type: array
|
||||
items:
|
||||
@@ -5959,7 +6036,14 @@ components:
|
||||
- input_tokens
|
||||
- output_tokens
|
||||
- total_tokens
|
||||
- cached_input_tokens
|
||||
- cache_creation_tokens
|
||||
- input_cost_usd
|
||||
- cached_input_cost_usd
|
||||
- cache_creation_cost_usd
|
||||
- output_cost_usd
|
||||
- cost_usd
|
||||
- cache_cost_usd
|
||||
- decision
|
||||
- entries
|
||||
AgentNetworkAccessLogSessionsResponse:
|
||||
@@ -6013,19 +6097,61 @@ components:
|
||||
total_tokens:
|
||||
type: integer
|
||||
format: int64
|
||||
description: Total tokens in the bucket.
|
||||
description: Total tokens in the bucket, including prompt-cache tokens.
|
||||
example: 184000
|
||||
cached_input_tokens:
|
||||
type: integer
|
||||
format: int64
|
||||
description: Total prompt-cache read tokens in the bucket.
|
||||
example: 20000
|
||||
cache_creation_tokens:
|
||||
type: integer
|
||||
format: int64
|
||||
description: Total prompt-cache write tokens in the bucket.
|
||||
example: 45000
|
||||
input_cost_usd:
|
||||
type: number
|
||||
format: double
|
||||
description: Total cost of non-cached input tokens in the bucket.
|
||||
example: 1.12
|
||||
cached_input_cost_usd:
|
||||
type: number
|
||||
format: double
|
||||
description: Total cost of prompt-cache read tokens in the bucket.
|
||||
example: 0.06
|
||||
cache_creation_cost_usd:
|
||||
type: number
|
||||
format: double
|
||||
description: Total cost of prompt-cache write tokens in the bucket.
|
||||
example: 0.36
|
||||
output_cost_usd:
|
||||
type: number
|
||||
format: double
|
||||
description: Total cost of output tokens in the bucket.
|
||||
example: 0.77
|
||||
cost_usd:
|
||||
type: number
|
||||
format: double
|
||||
description: Total estimated USD spend in the bucket.
|
||||
example: 2.31
|
||||
cache_cost_usd:
|
||||
type: number
|
||||
format: double
|
||||
description: Portion of cost_usd billed for prompt-cache usage in the bucket.
|
||||
example: 0.42
|
||||
required:
|
||||
- period_start
|
||||
- input_tokens
|
||||
- output_tokens
|
||||
- total_tokens
|
||||
- cached_input_tokens
|
||||
- cache_creation_tokens
|
||||
- input_cost_usd
|
||||
- cached_input_cost_usd
|
||||
- cache_creation_cost_usd
|
||||
- output_cost_usd
|
||||
- cost_usd
|
||||
- cache_cost_usd
|
||||
AgentNetworkSettings:
|
||||
type: object
|
||||
description: Per-account Agent Network gateway settings. One row per account; cluster and subdomain are auto-assigned on first provider create and immutable thereafter.
|
||||
|
||||
@@ -1732,6 +1732,21 @@ type AccountSettings struct {
|
||||
|
||||
// AgentNetworkAccessLog One per-request agent-network (LLM) access log entry with flattened, queryable LLM dimensions.
|
||||
type AgentNetworkAccessLog struct {
|
||||
// CacheCostUsd Portion of cost_usd billed for prompt-cache usage.
|
||||
CacheCostUsd float64 `json:"cache_cost_usd"`
|
||||
|
||||
// CacheCreationCostUsd Cost of the prompt-cache write tokens. Base component of cost_usd, and part of cache_cost_usd.
|
||||
CacheCreationCostUsd float64 `json:"cache_creation_cost_usd"`
|
||||
|
||||
// CacheCreationTokens Input tokens written to the provider's prompt cache. Zero for providers without a cache-write bucket.
|
||||
CacheCreationTokens int64 `json:"cache_creation_tokens"`
|
||||
|
||||
// CachedInputCostUsd Cost of the prompt-cache read tokens. Base component of cost_usd, and part of cache_cost_usd.
|
||||
CachedInputCostUsd float64 `json:"cached_input_cost_usd"`
|
||||
|
||||
// CachedInputTokens Input tokens read from the provider's prompt cache. Additive to input_tokens for Anthropic-shape providers; a subset of input_tokens for OpenAI.
|
||||
CachedInputTokens int64 `json:"cached_input_tokens"`
|
||||
|
||||
// CostUsd Estimated USD cost of the request.
|
||||
CostUsd float64 `json:"cost_usd"`
|
||||
|
||||
@@ -1753,6 +1768,9 @@ type AgentNetworkAccessLog struct {
|
||||
// Id Unique identifier for the access log entry.
|
||||
Id string `json:"id"`
|
||||
|
||||
// InputCostUsd Cost of the non-cached input tokens. Base component of cost_usd.
|
||||
InputCostUsd float64 `json:"input_cost_usd"`
|
||||
|
||||
// InputTokens Input (prompt) tokens consumed.
|
||||
InputTokens int64 `json:"input_tokens"`
|
||||
|
||||
@@ -1762,6 +1780,9 @@ type AgentNetworkAccessLog struct {
|
||||
// Model Requested LLM model.
|
||||
Model *string `json:"model,omitempty"`
|
||||
|
||||
// OutputCostUsd Cost of the output tokens. Base component of cost_usd.
|
||||
OutputCostUsd float64 `json:"output_cost_usd"`
|
||||
|
||||
// OutputTokens Output (completion) tokens produced.
|
||||
OutputTokens int64 `json:"output_tokens"`
|
||||
|
||||
@@ -1801,7 +1822,7 @@ type AgentNetworkAccessLog struct {
|
||||
// Timestamp Timestamp when the request was made.
|
||||
Timestamp time.Time `json:"timestamp"`
|
||||
|
||||
// TotalTokens Total tokens consumed.
|
||||
// TotalTokens Total tokens consumed, including prompt-cache tokens.
|
||||
TotalTokens int64 `json:"total_tokens"`
|
||||
|
||||
// UserId NetBird user id of the authenticated caller, if applicable.
|
||||
@@ -1810,6 +1831,21 @@ type AgentNetworkAccessLog struct {
|
||||
|
||||
// AgentNetworkAccessLogSession A session-grouped view of agent-network access logs — all requests sharing a session id (or a single session-less request) folded into one summary plus its ordered entries.
|
||||
type AgentNetworkAccessLogSession struct {
|
||||
// CacheCostUsd Portion of cost_usd billed for prompt-cache usage across the session.
|
||||
CacheCostUsd float64 `json:"cache_cost_usd"`
|
||||
|
||||
// CacheCreationCostUsd Total cost of prompt-cache write tokens across the session.
|
||||
CacheCreationCostUsd float64 `json:"cache_creation_cost_usd"`
|
||||
|
||||
// CacheCreationTokens Total prompt-cache write tokens across the session.
|
||||
CacheCreationTokens int64 `json:"cache_creation_tokens"`
|
||||
|
||||
// CachedInputCostUsd Total cost of prompt-cache read tokens across the session.
|
||||
CachedInputCostUsd float64 `json:"cached_input_cost_usd"`
|
||||
|
||||
// CachedInputTokens Total prompt-cache read tokens across the session.
|
||||
CachedInputTokens int64 `json:"cached_input_tokens"`
|
||||
|
||||
// CostUsd Total estimated USD cost across the session.
|
||||
CostUsd float64 `json:"cost_usd"`
|
||||
|
||||
@@ -1825,12 +1861,18 @@ type AgentNetworkAccessLogSession struct {
|
||||
// GroupIds Union of the authorising group ids across the session's entries.
|
||||
GroupIds *[]string `json:"group_ids,omitempty"`
|
||||
|
||||
// InputCostUsd Total cost of non-cached input tokens across the session.
|
||||
InputCostUsd float64 `json:"input_cost_usd"`
|
||||
|
||||
// InputTokens Total input (prompt) tokens across the session.
|
||||
InputTokens int64 `json:"input_tokens"`
|
||||
|
||||
// Models Distinct models seen in the session.
|
||||
Models *[]string `json:"models,omitempty"`
|
||||
|
||||
// OutputCostUsd Total cost of output tokens across the session.
|
||||
OutputCostUsd float64 `json:"output_cost_usd"`
|
||||
|
||||
// OutputTokens Total output (completion) tokens across the session.
|
||||
OutputTokens int64 `json:"output_tokens"`
|
||||
|
||||
@@ -1846,7 +1888,7 @@ type AgentNetworkAccessLogSession struct {
|
||||
// StartedAt Timestamp of the session's earliest request.
|
||||
StartedAt time.Time `json:"started_at"`
|
||||
|
||||
// TotalTokens Total tokens across the session.
|
||||
// TotalTokens Total tokens across the session, including prompt-cache tokens.
|
||||
TotalTokens int64 `json:"total_tokens"`
|
||||
|
||||
// UserId NetBird user id of the session's caller.
|
||||
@@ -2347,19 +2389,40 @@ type AgentNetworkSettingsRequest struct {
|
||||
|
||||
// AgentNetworkUsageBucket One aggregated agent-network usage time bucket (UTC). The bucket width is set by the request's granularity.
|
||||
type AgentNetworkUsageBucket struct {
|
||||
// CacheCostUsd Portion of cost_usd billed for prompt-cache usage in the bucket.
|
||||
CacheCostUsd float64 `json:"cache_cost_usd"`
|
||||
|
||||
// CacheCreationCostUsd Total cost of prompt-cache write tokens in the bucket.
|
||||
CacheCreationCostUsd float64 `json:"cache_creation_cost_usd"`
|
||||
|
||||
// CacheCreationTokens Total prompt-cache write tokens in the bucket.
|
||||
CacheCreationTokens int64 `json:"cache_creation_tokens"`
|
||||
|
||||
// CachedInputCostUsd Total cost of prompt-cache read tokens in the bucket.
|
||||
CachedInputCostUsd float64 `json:"cached_input_cost_usd"`
|
||||
|
||||
// CachedInputTokens Total prompt-cache read tokens in the bucket.
|
||||
CachedInputTokens int64 `json:"cached_input_tokens"`
|
||||
|
||||
// CostUsd Total estimated USD spend in the bucket.
|
||||
CostUsd float64 `json:"cost_usd"`
|
||||
|
||||
// InputCostUsd Total cost of non-cached input tokens in the bucket.
|
||||
InputCostUsd float64 `json:"input_cost_usd"`
|
||||
|
||||
// InputTokens Total input (prompt) tokens in the bucket.
|
||||
InputTokens int64 `json:"input_tokens"`
|
||||
|
||||
// OutputCostUsd Total cost of output tokens in the bucket.
|
||||
OutputCostUsd float64 `json:"output_cost_usd"`
|
||||
|
||||
// OutputTokens Total output (completion) tokens in the bucket.
|
||||
OutputTokens int64 `json:"output_tokens"`
|
||||
|
||||
// PeriodStart Start of the bucket in YYYY-MM-DD (UTC) — the day, the week start (Monday), or the month start, depending on granularity.
|
||||
PeriodStart string `json:"period_start"`
|
||||
|
||||
// TotalTokens Total tokens in the bucket.
|
||||
// TotalTokens Total tokens in the bucket, including prompt-cache tokens.
|
||||
TotalTokens int64 `json:"total_tokens"`
|
||||
}
|
||||
|
||||
|
||||
Reference in New Issue
Block a user