api

package
v1.0.2 Latest Latest
Warning

This package is not in the latest version of its module.

Go to latest
Published: Oct 11, 2026 License: Apache-2.0 Imports: 25 Imported by: 0

Documentation

Index

Constants

This section is empty.

Variables

This section is empty.

Functions

This section is empty.

Types

type APIKeyUsage added in v0.9.11

type APIKeyUsage struct {
	Service UsageQuerier
}

func (*APIKeyUsage) Query added in v0.9.11

func (a *APIKeyUsage) Query(c *gin.Context)

Query returns the usage ranking only to the authenticated Root principal. @Tags DashboardAPI @Security ApiKeyAuth @Summary Query Root-only client API key usage ranking @Param time_range query string false "today, 1h, 6h, 24h, 7d" @Param sort_by query string false "tokens, request_count, cost" @Param limit query int false "10, 20, 50" @Success 200 {object} util.ResponseResult{data=schema.Response} @Failure 400,401,403,503 {object} util.ResponseResult @Router /api/v1/dashboard/api-key-ranking [get]

type CircuitBreakerInfo

type CircuitBreakerInfo struct {
	ID           string `json:"id"`
	Type         string `json:"type"`          // "endpoint" 或 "service"
	Name         string `json:"name"`          // 显示名称
	ModelID      string `json:"model_id"`      // 关联模型 ID
	ModelName    string `json:"model_name"`    // 关联模型名称
	ProviderID   string `json:"provider_id"`   // 关联供应商 ID
	ProviderName string `json:"provider_name"` // 关联供应商名称
	URL          string `json:"url"`           // 关联的 URL 地址
}

type Dashboard

type Dashboard struct {
	DB          *gorm.DB
	RedisClient *redis.Client
	RedisSync   *biz.ConfigRedisSync
	// contains filtered or unexported fields
}

func (*Dashboard) HandleWebSocket

func (a *Dashboard) HandleWebSocket(c *gin.Context)

HandleWebSocket upgrades the HTTP request to WebSocket and pushes dashboard metrics

func (*Dashboard) IsPrometheusAvailable

func (a *Dashboard) IsPrometheusAvailable() bool

IsPrometheusAvailable 导出的版本,用于测试

func (*Dashboard) QueryCircuitBreakers

func (a *Dashboard) QueryCircuitBreakers(c *gin.Context)

@Tags DashboardAPI @Security ApiKeyAuth @Summary Query all endpoints that are currently in Open/Isolated state from Redis @Success 200 {object} util.ResponseResult{data=[]CircuitBreakerInfo} @Router /api/v1/dashboard/circuit-breakers [get]

func (*Dashboard) QueryModelPerformanceTrends added in v0.9.3

func (a *Dashboard) QueryModelPerformanceTrends(c *gin.Context)

@Tags DashboardAPI @Security ApiKeyAuth @Summary Query model TTFT and request-period OTPS trends @Param model query string true "Model code" @Param time_range query string false "Time range: 1h, 6h, 24h, 7d, today, or a duration up to 7d (default: 1h)" @Param end_time query string false "RFC3339 query end time" @Success 200 {object} util.ResponseResult{data=ModelPerformanceTrendsResponse} @Failure 400 {object} util.ResponseResult @Router /api/v1/dashboard/model-performance-trends [get]

func (*Dashboard) QueryModelRanking

func (a *Dashboard) QueryModelRanking(c *gin.Context)

@Tags DashboardAPI @Security ApiKeyAuth @Summary Query model usage ranking with detailed metrics @Param sort_by query string false "Sort by: request_count, avg_latency, avg_ttft, tokens, cost, success_rate, otps (default: request_count)" @Param limit query int false "Limit results (default: 10)" @Param model query string false "Filter by model id or model code" @Success 200 {object} util.ResponseResult{data=[]ModelRankingItem} @Router /api/v1/dashboard/model-ranking [get]

func (*Dashboard) QueryOverview

func (a *Dashboard) QueryOverview(c *gin.Context)

@Tags DashboardAPI @Security ApiKeyAuth @Summary Query dashboard overview metrics (QPS, daily stats, latency, circuit breakers) @Success 200 {object} util.ResponseResult{data=OverviewResponse} @Router /api/v1/dashboard/overview [get]

func (*Dashboard) QueryProviderRanking added in v0.9.10

func (a *Dashboard) QueryProviderRanking(c *gin.Context)

@Tags DashboardAPI @Security ApiKeyAuth @Summary Query provider upstream health ranking @Param sort_by query string false "Sort by: request_count, avg_latency, avg_ttft, tokens, cost, success_rate, otps (default: request_count)" @Param time_range query string false "Time range: 1h, 6h, 24h, 7d, today (max 7d, default: 1h)" @Param limit query int false "Limit results (default: 10)" @Param provider query string false "Filter by provider id or code" @Param end_time query string false "RFC3339 query end time" @Success 200 {object} util.ResponseResult{data=[]ProviderRankingItem} @Failure 500 {object} util.ResponseResult @Router /api/v1/dashboard/provider-ranking [get]

func (*Dashboard) QueryQPS

func (a *Dashboard) QueryQPS(c *gin.Context)

@Tags DashboardAPI @Security ApiKeyAuth @Summary Query real-time gateway QPS from Prometheus @Success 200 {object} util.ResponseResult{data=float64} @Router /api/v1/dashboard/qps [get]

func (*Dashboard) QueryTrends

func (a *Dashboard) QueryTrends(c *gin.Context)

@Tags DashboardAPI @Security ApiKeyAuth @Summary Query bucketed gateway traffic success/failure trends @Param group_by query string false "Group by: model, provider, tenant, endpoint (default: global)" @Param time_range query string false "Time range: 1h, 6h, 24h, 7d, today (default: 1h)" @Param model query string false "Filter by model code" @Success 200 {object} util.ResponseResult{data=TrendsResponse} @Router /api/v1/dashboard/trends [get]

func (*Dashboard) SyncRedis

func (a *Dashboard) SyncRedis(c *gin.Context)

@Tags DashboardAPI @Security ApiKeyAuth @Summary One-click synchronization of database configuration to Redis @Success 200 {object} util.ResponseResult @Router /api/v1/dashboard/sync-redis [post]

type ModelPerformanceTrendsResponse added in v0.9.3

type ModelPerformanceTrendsResponse struct {
	Times     []string   `json:"times"`
	AvgTTFTMs []*float64 `json:"avg_ttft_ms"`
	Otps      []*float64 `json:"otps"`
}

type ModelRankingItem

type ModelRankingItem struct {
	ModelID      string  `json:"model_id"`
	ModelCode    string  `json:"model_code"`
	ModelName    string  `json:"model_name"`
	RequestCount int64   `json:"request_count"`
	SuccessCount int64   `json:"success_count"`
	FailCount    int64   `json:"fail_count"`
	SuccessRate  float64 `json:"success_rate"`
	AvgLatencyMs float64 `json:"avg_latency_ms"`
	P50LatencyMs float64 `json:"p50_latency_ms"`
	P95LatencyMs float64 `json:"p95_latency_ms"`
	P99LatencyMs float64 `json:"p99_latency_ms"`
	AvgTTFTMs    float64 `json:"avg_ttft_ms"`
	P50TTFTMs    float64 `json:"p50_ttft_ms"`
	P95TTFTMs    float64 `json:"p95_ttft_ms"`
	P99TTFTMs    float64 `json:"p99_ttft_ms"`
	TotalTokens  int64   `json:"total_tokens"`
	TotalCost    float64 `json:"total_cost"`
	Otps         float64 `json:"otps"`
}

type OverviewResponse

type OverviewResponse struct {
	QPS                      float64              `json:"qps"`
	DailyRequests            int64                `json:"daily_requests"`
	DailyPromptTokens        int64                `json:"daily_prompt_tokens"`
	DailyCompletionTokens    int64                `json:"daily_completion_tokens"`
	DailyCachedTokens        int64                `json:"daily_cached_tokens"`
	DailyCacheCreationTokens int64                `json:"daily_cache_creation_tokens"`
	DailyCost                float64              `json:"daily_cost"`
	AvgLatencyMs             float64              `json:"avg_latency_ms"`
	AvgTTFTMs                float64              `json:"avg_ttft_ms"`
	ActiveCircuitBreakers    []CircuitBreakerInfo `json:"active_circuit_breakers"`
}

type ProviderRankingItem added in v0.9.10

type ProviderRankingItem struct {
	ProviderID   string `json:"provider_id"`
	ProviderCode string `json:"provider_code"`
	ProviderName string `json:"provider_name"`
	// UpstreamCallCount 上游调用次数。故障转移产生的多次尝试各自计入对应供应商。
	UpstreamCallCount int64 `json:"upstream_call_count"`
	SuccessCount      int64 `json:"success_count"`
	FailCount         int64 `json:"fail_count"`
	// UpstreamCallSuccessRate 上游调用成功率(百分比)。
	// 与 Request Success Rate 分母不同:故障转移走掉的失败尝试计入此处的失败数。
	UpstreamCallSuccessRate float64 `json:"upstream_call_success_rate"`
	AvgLatencyMs            float64 `json:"avg_latency_ms"`
	AvgTTFTMs               float64 `json:"avg_ttft_ms"`
	TotalTokens             int64   `json:"total_tokens"`
	TotalCost               float64 `json:"total_cost"`
	Otps                    float64 `json:"otps"`
	EndpointCount           int64   `json:"endpoint_count"`
	OpenBreakerCount        int64   `json:"open_breaker_count"`
}

ProviderRankingItem 单个供应商在给定时间窗口内的聚合指标。

注意:不含 P50/P95/P99 分位数。内存 store 只累积 sum/count 而无直方图桶, 分位数仅 Prometheus 可算;一个在部分部署下静默消失的指标不如不提供。详见 ADR-0008。

type TrendsResponse

type TrendsResponse struct {
	Times  []string       `json:"times"`
	Series []TrendsSeries `json:"series"`
}

type TrendsSeries

type TrendsSeries struct {
	Label   string  `json:"label"`
	Success []int64 `json:"success"`
	Failure []int64 `json:"failure"`
	Total   []int64 `json:"total"`
}

type UsageQuerier added in v0.9.11

type UsageQuerier interface {
	Query(context.Context, schema.Query) (*schema.Response, error)
}

type WSConfigMsg

type WSConfigMsg struct {
	TrendsTimeRange       string `json:"trends_time_range"`
	TrendsGroupBy         string `json:"trends_group_by"`
	ModelRankingSortBy    string `json:"model_ranking_sort_by"`
	ModelRankingTimeRange string `json:"model_ranking_time_range"`
}

WSConfigMsg represents the configuration parameters sent by the client

type WSDashboardPayload

type WSDashboardPayload struct {
	Overview     *OverviewResponse  `json:"overview"`
	Trends       *TrendsResponse    `json:"trends"`
	ModelRanking []ModelRankingItem `json:"model_ranking"`
}

WSDashboardPayload represents the unified push data payload

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL