Skip to content

GitLab

Commit 90bce60b authored by yangjianbo's avatar yangjianbo
Browse files

feat: merge dev

parent a458e684
Show whitespace changes
Inline Side-by-side
backend/internal/service/ops_service.go
View file @ 90bce60b
......@@ -208,6 +208,25 @@ func (s *OpsService) RecordError(ctx context.Context, entry *OpsInsertErrorLogIn
out.Detail = ""
}
out.UpstreamRequestBody = strings.TrimSpace(out.UpstreamRequestBody)
if out.UpstreamRequestBody != "" {
// Reuse the same sanitization/trimming strategy as request body storage.
// Keep it small so it is safe to persist in ops_error_logs JSON.
sanitized, truncated, _ := sanitizeAndTrimRequestBody([]byte(out.UpstreamRequestBody), 10*1024)
if sanitized != "" {
out.UpstreamRequestBody = sanitized
if truncated {
out.Kind = strings.TrimSpace(out.Kind)
if out.Kind == "" {
out.Kind = "upstream"
}
out.Kind = out.Kind + ":request_body_truncated"
}
} else {
out.UpstreamRequestBody = ""
}
}
// Drop fully-empty events (can happen if only status code was known).
if out.UpstreamStatusCode == 0 && out.Message == "" && out.Detail == "" {
continue
......@@ -236,7 +255,13 @@ func (s *OpsService) GetErrorLogs(ctx context.Context, filter *OpsErrorLogFilter
if s.opsRepo == nil {
return &OpsErrorLogList{Errors: []*OpsErrorLog{}, Total: 0, Page: 1, PageSize: 20}, nil
}
return s.opsRepo.ListErrorLogs(ctx, filter)
result, err := s.opsRepo.ListErrorLogs(ctx, filter)
if err != nil {
log.Printf("[Ops] GetErrorLogs failed: %v", err)
return nil, err
}
return result, nil
}
func (s *OpsService) GetErrorLogByID(ctx context.Context, id int64) (*OpsErrorLogDetail, error) {
......@@ -256,6 +281,46 @@ func (s *OpsService) GetErrorLogByID(ctx context.Context, id int64) (*OpsErrorLo
return detail, nil
}
func (s *OpsService) ListRetryAttemptsByErrorID(ctx context.Context, errorID int64, limit int) ([]*OpsRetryAttempt, error) {
if err := s.RequireMonitoringEnabled(ctx); err != nil {
return nil, err
}
if s.opsRepo == nil {
return nil, infraerrors.ServiceUnavailable("OPS_REPO_UNAVAILABLE", "Ops repository not available")
}
if errorID <= 0 {
return nil, infraerrors.BadRequest("OPS_ERROR_INVALID_ID", "invalid error id")
}
items, err := s.opsRepo.ListRetryAttemptsByErrorID(ctx, errorID, limit)
if err != nil {
if errors.Is(err, sql.ErrNoRows) {
return []*OpsRetryAttempt{}, nil
}
return nil, infraerrors.InternalServer("OPS_RETRY_LIST_FAILED", "Failed to list retry attempts").WithCause(err)
}
return items, nil
}
func (s *OpsService) UpdateErrorResolution(ctx context.Context, errorID int64, resolved bool, resolvedByUserID *int64, resolvedRetryID *int64) error {
if err := s.RequireMonitoringEnabled(ctx); err != nil {
return err
}
if s.opsRepo == nil {
return infraerrors.ServiceUnavailable("OPS_REPO_UNAVAILABLE", "Ops repository not available")
}
if errorID <= 0 {
return infraerrors.BadRequest("OPS_ERROR_INVALID_ID", "invalid error id")
}
// Best-effort ensure the error exists
if _, err := s.opsRepo.GetErrorLogByID(ctx, errorID); err != nil {
if errors.Is(err, sql.ErrNoRows) {
return infraerrors.NotFound("OPS_ERROR_NOT_FOUND", "ops error log not found")
}
return infraerrors.InternalServer("OPS_ERROR_LOAD_FAILED", "Failed to load ops error log").WithCause(err)
}
return s.opsRepo.UpdateErrorResolution(ctx, errorID, resolved, resolvedByUserID, resolvedRetryID, nil)
}
func sanitizeAndTrimRequestBody(raw []byte, maxBytes int) (jsonString string, truncated bool, bytesLen int) {
bytesLen = len(raw)
if len(raw) == 0 {
......@@ -296,14 +361,34 @@ func sanitizeAndTrimRequestBody(raw []byte, maxBytes int) (jsonString string, tr
}
}
// Last resort: store a minimal placeholder (still valid JSON).
placeholder := map[string]any{
"request_body_truncated": true,
// Last resort: keep JSON shape but drop big fields.
// This avoids downstream code that expects certain top-level keys from crashing.
if root, ok := decoded.(map[string]any); ok {
placeholder := shallowCopyMap(root)
placeholder["request_body_truncated"] = true
// Replace potentially huge arrays/strings, but keep the keys present.
for _, k := range []string{"messages", "contents", "input", "prompt"} {
if _, exists := placeholder[k]; exists {
placeholder[k] = []any{}
}
if model := extractString(decoded, "model"); model != "" {
placeholder["model"] = model
}
for _, k := range []string{"text"} {
if _, exists := placeholder[k]; exists {
placeholder[k] = ""
}
}
encoded4, err4 := json.Marshal(placeholder)
if err4 == nil {
if len(encoded4) <= maxBytes {
return string(encoded4), true, bytesLen
}
}
}
// Final fallback: minimal valid JSON.
encoded4, err4 := json.Marshal(map[string]any{"request_body_truncated": true})
if err4 != nil {
return "", true, bytesLen
}
......@@ -526,12 +611,3 @@ func sanitizeErrorBodyForStorage(raw string, maxBytes int) (sanitized string, tr
}
return raw, false
}
func extractString(v any, key string) string {
root, ok := v.(map[string]any)
if !ok {
return ""
}
s, _ := root[key].(string)
return strings.TrimSpace(s)
}
backend/internal/service/ops_settings.go
View file @ 90bce60b
......@@ -369,6 +369,8 @@ func defaultOpsAdvancedSettings() *OpsAdvancedSettings {
AggregationEnabled: false,
},
IgnoreCountTokensErrors: false,
IgnoreContextCanceled: true, // Default to true - client disconnects are not errors
IgnoreNoAvailableAccounts: false, // Default to false - this is a real routing issue
AutoRefreshEnabled: false,
AutoRefreshIntervalSec: 30,
}
......@@ -482,13 +484,11 @@ const SettingKeyOpsMetricThresholds = "ops_metric_thresholds"
func defaultOpsMetricThresholds() *OpsMetricThresholds {
slaMin := 99.5
latencyMax := 2000.0
ttftMax := 500.0
reqErrMax := 5.0
upstreamErrMax := 5.0
return &OpsMetricThresholds{
SLAPercentMin: &slaMin,
LatencyP99MsMax: &latencyMax,
TTFTp99MsMax: &ttftMax,
RequestErrorRatePercentMax: &reqErrMax,
UpstreamErrorRatePercentMax: &upstreamErrMax,
......@@ -538,9 +538,6 @@ func (s *OpsService) UpdateMetricThresholds(ctx context.Context, cfg *OpsMetricT
if cfg.SLAPercentMin != nil && (*cfg.SLAPercentMin < 0 || *cfg.SLAPercentMin > 100) {
return nil, errors.New("sla_percent_min must be between 0 and 100")
}
if cfg.LatencyP99MsMax != nil && *cfg.LatencyP99MsMax < 0 {
return nil, errors.New("latency_p99_ms_max must be >= 0")
}
if cfg.TTFTp99MsMax != nil && *cfg.TTFTp99MsMax < 0 {
return nil, errors.New("ttft_p99_ms_max must be >= 0")
}
......
backend/internal/service/ops_settings_models.go
View file @ 90bce60b
......@@ -63,7 +63,6 @@ type OpsAlertSilencingSettings struct {
type OpsMetricThresholds struct {
SLAPercentMin *float64 `json:"sla_percent_min,omitempty"` // SLA低于此值变红
LatencyP99MsMax *float64 `json:"latency_p99_ms_max,omitempty"` // 延迟P99高于此值变红
TTFTp99MsMax *float64 `json:"ttft_p99_ms_max,omitempty"` // TTFT P99高于此值变红
RequestErrorRatePercentMax *float64 `json:"request_error_rate_percent_max,omitempty"` // 请求错误率高于此值变红
UpstreamErrorRatePercentMax *float64 `json:"upstream_error_rate_percent_max,omitempty"` // 上游错误率高于此值变红
......@@ -82,6 +81,8 @@ type OpsAdvancedSettings struct {
DataRetention OpsDataRetentionSettings `json:"data_retention"`
Aggregation OpsAggregationSettings `json:"aggregation"`
IgnoreCountTokensErrors bool `json:"ignore_count_tokens_errors"`
IgnoreContextCanceled bool `json:"ignore_context_canceled"`
IgnoreNoAvailableAccounts bool `json:"ignore_no_available_accounts"`
AutoRefreshEnabled bool `json:"auto_refresh_enabled"`
AutoRefreshIntervalSec int `json:"auto_refresh_interval_seconds"`
}
......
backend/internal/service/ops_upstream_context.go
View file @ 90bce60b
......@@ -15,6 +15,11 @@ const (
OpsUpstreamErrorMessageKey = "ops_upstream_error_message"
OpsUpstreamErrorDetailKey = "ops_upstream_error_detail"
OpsUpstreamErrorsKey = "ops_upstream_errors"
// Best-effort capture of the current upstream request body so ops can
// retry the specific upstream attempt (not just the client request).
// This value is sanitized+trimmed before being persisted.
OpsUpstreamRequestBodyKey = "ops_upstream_request_body"
)
func setOpsUpstreamError(c *gin.Context, upstreamStatusCode int, upstreamMessage, upstreamDetail string) {
......@@ -40,11 +45,19 @@ type OpsUpstreamErrorEvent struct {
// Context
Platform string `json:"platform,omitempty"`
AccountID int64 `json:"account_id,omitempty"`
AccountName string `json:"account_name,omitempty"`
// Outcome
UpstreamStatusCode int `json:"upstream_status_code,omitempty"`
UpstreamRequestID string `json:"upstream_request_id,omitempty"`
// Best-effort upstream request capture (sanitized+trimmed).
// Required for retrying a specific upstream attempt.
UpstreamRequestBody string `json:"upstream_request_body,omitempty"`
// Best-effort upstream response capture (sanitized+trimmed).
UpstreamResponseBody string `json:"upstream_response_body,omitempty"`
// Kind: http_error | request_error | retry_exhausted | failover
Kind string `json:"kind,omitempty"`
......@@ -61,6 +74,8 @@ func appendOpsUpstreamError(c *gin.Context, ev OpsUpstreamErrorEvent) {
}
ev.Platform = strings.TrimSpace(ev.Platform)
ev.UpstreamRequestID = strings.TrimSpace(ev.UpstreamRequestID)
ev.UpstreamRequestBody = strings.TrimSpace(ev.UpstreamRequestBody)
ev.UpstreamResponseBody = strings.TrimSpace(ev.UpstreamResponseBody)
ev.Kind = strings.TrimSpace(ev.Kind)
ev.Message = strings.TrimSpace(ev.Message)
ev.Detail = strings.TrimSpace(ev.Detail)
......@@ -68,6 +83,16 @@ func appendOpsUpstreamError(c *gin.Context, ev OpsUpstreamErrorEvent) {
ev.Message = sanitizeUpstreamErrorMessage(ev.Message)
}
// If the caller didn't explicitly pass upstream request body but the gateway
// stored it on the context, attach it so ops can retry this specific attempt.
if ev.UpstreamRequestBody == "" {
if v, ok := c.Get(OpsUpstreamRequestBodyKey); ok {
if s, ok := v.(string); ok {
ev.UpstreamRequestBody = strings.TrimSpace(s)
}
}
}
var existing []*OpsUpstreamErrorEvent
if v, ok := c.Get(OpsUpstreamErrorsKey); ok {
if arr, ok := v.([]*OpsUpstreamErrorEvent); ok {
......@@ -92,3 +117,15 @@ func marshalOpsUpstreamErrors(events []*OpsUpstreamErrorEvent) *string {
s := string(raw)
return &s
}
func ParseOpsUpstreamErrors(raw string) ([]*OpsUpstreamErrorEvent, error) {
raw = strings.TrimSpace(raw)
if raw == "" {
return []*OpsUpstreamErrorEvent{}, nil
}
var out []*OpsUpstreamErrorEvent
if err := json.Unmarshal([]byte(raw), &out); err != nil {
return nil, err
}
return out, nil
}
backend/internal/service/proxy.go
View file @ 90bce60b
......@@ -32,4 +32,15 @@ func (p *Proxy) URL() string {
type ProxyWithAccountCount struct {
Proxy
AccountCount int64
LatencyMs *int64
LatencyStatus string
LatencyMessage string
}
type ProxyAccountSummary struct {
ID int64
Name string
Platform string
Type string
Notes *string
}
backend/internal/service/proxy_latency_cache.go 0 → 100644
View file @ 90bce60b
package service
import (
"context"
"time"
)
type ProxyLatencyInfo struct {
Success bool `json:"success"`
LatencyMs *int64 `json:"latency_ms,omitempty"`
Message string `json:"message,omitempty"`
UpdatedAt time.Time `json:"updated_at"`
}
type ProxyLatencyCache interface {
GetProxyLatencies(ctx context.Context, proxyIDs []int64) (map[int64]*ProxyLatencyInfo, error)
SetProxyLatency(ctx context.Context, proxyID int64, info *ProxyLatencyInfo) error
}
backend/internal/service/proxy_service.go
View file @ 90bce60b
......@@ -10,6 +10,7 @@ import (
var (
ErrProxyNotFound = infraerrors.NotFound("PROXY_NOT_FOUND", "proxy not found")
ErrProxyInUse = infraerrors.Conflict("PROXY_IN_USE", "proxy is in use by accounts")
)
type ProxyRepository interface {
......@@ -26,6 +27,7 @@ type ProxyRepository interface {
ExistsByHostPortAuth(ctx context.Context, host string, port int, username, password string) (bool, error)
CountAccountsByProxyID(ctx context.Context, proxyID int64) (int64, error)
ListAccountSummariesByProxyID(ctx context.Context, proxyID int64) ([]ProxyAccountSummary, error)
}
// CreateProxyRequest 创建代理请求
......
backend/internal/service/ratelimit_service.go
View file @ 90bce60b
......@@ -179,7 +179,7 @@ func (s *RateLimitService) PreCheckUsage(ctx context.Context, account *Account,
start := geminiDailyWindowStart(now)
totals, ok := s.getGeminiUsageTotals(account.ID, start, now)
if !ok {
stats, err := s.usageRepo.GetModelStatsWithFilters(ctx, start, now, 0, 0, account.ID)
stats, err := s.usageRepo.GetModelStatsWithFilters(ctx, start, now, 0, 0, account.ID, 0, nil)
if err != nil {
return true, err
}
......@@ -226,7 +226,7 @@ func (s *RateLimitService) PreCheckUsage(ctx context.Context, account *Account,
if limit > 0 {
start := now.Truncate(time.Minute)
stats, err := s.usageRepo.GetModelStatsWithFilters(ctx, start, now, 0, 0, account.ID)
stats, err := s.usageRepo.GetModelStatsWithFilters(ctx, start, now, 0, 0, account.ID, 0, nil)
if err != nil {
return true, err
}
......
backend/internal/service/usage_log.go
View file @ 90bce60b
......@@ -33,6 +33,8 @@ type UsageLog struct {
TotalCost float64
ActualCost float64
RateMultiplier float64
// AccountRateMultiplier 账号计费倍率快照（nil 表示历史数据，按 1.0 处理）
AccountRateMultiplier *float64
BillingType int8
Stream bool
......
backend/migrations/037_add_account_rate_multiplier.sql 0 → 100644
View file @ 90bce60b
-- Add account billing rate multiplier and per-usage snapshot.
--
-- accounts.rate_multiplier: 账号计费倍率（>=0，允许 0 表示该账号计费为 0）。
-- usage_logs.account_rate_multiplier: 每条 usage log 的账号倍率快照，用于实现
-- “倍率调整仅影响之后请求”，并支持同一天分段倍率加权统计。
--
-- 注意：usage_logs.account_rate_multiplier 不做回填、不设置 NOT NULL。
-- 老数据为 NULL 时，统计口径按 1.0 处理（COALESCE）。
ALTER TABLE IF EXISTS accounts
ADD COLUMN IF NOT EXISTS rate_multiplier DECIMAL(10,4) NOT NULL DEFAULT 1.0;
ALTER TABLE IF EXISTS usage_logs
ADD COLUMN IF NOT EXISTS account_rate_multiplier DECIMAL(10,4);
backend/migrations/037_ops_alert_silences.sql 0 → 100644
View file @ 90bce60b
-- +goose Up
-- +goose StatementBegin
-- Ops alert silences: scoped (rule_id + platform + group_id + region)
CREATE TABLE IF NOT EXISTS ops_alert_silences (
id BIGSERIAL PRIMARY KEY,
rule_id BIGINT NOT NULL,
platform VARCHAR(64) NOT NULL,
group_id BIGINT,
region VARCHAR(64),
until TIMESTAMPTZ NOT NULL,
reason TEXT,
created_by BIGINT,
created_at TIMESTAMPTZ NOT NULL DEFAULT NOW()
);
CREATE INDEX IF NOT EXISTS idx_ops_alert_silences_lookup
ON ops_alert_silences (rule_id, platform, group_id, region, until);
-- +goose StatementEnd
-- +goose Down
-- +goose StatementBegin
DROP TABLE IF EXISTS ops_alert_silences;
-- +goose StatementEnd
backend/migrations/038_ops_errors_resolution_retry_results_and_standardize_classification.sql 0 → 100644
View file @ 90bce60b
-- Add resolution tracking to ops_error_logs, persist retry results, and standardize error classification enums.
--
-- This migration is intentionally idempotent.
SET LOCAL lock_timeout = '5s';
SET LOCAL statement_timeout = '10min';
-- ============================================
-- 1) ops_error_logs: resolution fields
-- ============================================
ALTER TABLE ops_error_logs
ADD COLUMN IF NOT EXISTS resolved BOOLEAN NOT NULL DEFAULT false;
ALTER TABLE ops_error_logs
ADD COLUMN IF NOT EXISTS resolved_at TIMESTAMPTZ;
ALTER TABLE ops_error_logs
ADD COLUMN IF NOT EXISTS resolved_by_user_id BIGINT;
ALTER TABLE ops_error_logs
ADD COLUMN IF NOT EXISTS resolved_retry_id BIGINT;
CREATE INDEX IF NOT EXISTS idx_ops_error_logs_resolved_time
ON ops_error_logs (resolved, created_at DESC);
CREATE INDEX IF NOT EXISTS idx_ops_error_logs_unresolved_time
ON ops_error_logs (created_at DESC)
WHERE resolved = false;
-- ============================================
-- 2) ops_retry_attempts: persist execution results
-- ============================================
ALTER TABLE ops_retry_attempts
ADD COLUMN IF NOT EXISTS success BOOLEAN;
ALTER TABLE ops_retry_attempts
ADD COLUMN IF NOT EXISTS http_status_code INT;
ALTER TABLE ops_retry_attempts
ADD COLUMN IF NOT EXISTS upstream_request_id VARCHAR(128);
ALTER TABLE ops_retry_attempts
ADD COLUMN IF NOT EXISTS used_account_id BIGINT;
ALTER TABLE ops_retry_attempts
ADD COLUMN IF NOT EXISTS response_preview TEXT;
ALTER TABLE ops_retry_attempts
ADD COLUMN IF NOT EXISTS response_truncated BOOLEAN NOT NULL DEFAULT false;
CREATE INDEX IF NOT EXISTS idx_ops_retry_attempts_success_time
ON ops_retry_attempts (success, created_at DESC);
-- Backfill best-effort fields for existing rows.
UPDATE ops_retry_attempts
SET success = (LOWER(COALESCE(status, '')) = 'succeeded')
WHERE success IS NULL;
UPDATE ops_retry_attempts
SET upstream_request_id = result_request_id
WHERE upstream_request_id IS NULL AND result_request_id IS NOT NULL;
-- ============================================
-- 3) Standardize classification enums in ops_error_logs
--
-- New enums:
-- error_phase: request|auth|routing|upstream|network|internal
-- error_owner: client|provider|platform
-- error_source: client_request|upstream_http|gateway
-- ============================================
-- Owner: legacy sub2api => platform.
UPDATE ops_error_logs
SET error_owner = 'platform'
WHERE LOWER(COALESCE(error_owner, '')) = 'sub2api';
-- Owner: normalize empty/null to platform (best-effort).
UPDATE ops_error_logs
SET error_owner = 'platform'
WHERE COALESCE(TRIM(error_owner), '') = '';
-- Phase: map legacy phases.
UPDATE ops_error_logs
SET error_phase = CASE
WHEN COALESCE(TRIM(error_phase), '') = '' THEN 'internal'
WHEN LOWER(error_phase) IN ('billing', 'concurrency', 'response') THEN 'request'
WHEN LOWER(error_phase) IN ('scheduling') THEN 'routing'
WHEN LOWER(error_phase) IN ('request', 'auth', 'routing', 'upstream', 'network', 'internal') THEN LOWER(error_phase)
ELSE 'internal'
END;
-- Source: map legacy sources.
UPDATE ops_error_logs
SET error_source = CASE
WHEN COALESCE(TRIM(error_source), '') = '' THEN 'gateway'
WHEN LOWER(error_source) IN ('billing', 'concurrency') THEN 'client_request'
WHEN LOWER(error_source) IN ('upstream_http') THEN 'upstream_http'
WHEN LOWER(error_source) IN ('upstream_network') THEN 'gateway'
WHEN LOWER(error_source) IN ('internal') THEN 'gateway'
WHEN LOWER(error_source) IN ('client_request', 'upstream_http', 'gateway') THEN LOWER(error_source)
ELSE 'gateway'
END;
-- Auto-resolve recovered upstream errors (client status < 400).
UPDATE ops_error_logs
SET
resolved = true,
resolved_at = COALESCE(resolved_at, created_at)
WHERE resolved = false AND COALESCE(status_code, 0) > 0 AND COALESCE(status_code, 0) < 400;
frontend/src/api/admin/dashboard.ts
View file @ 90bce60b
......@@ -46,6 +46,10 @@ export interface TrendParams {
granularity?: 'day' | 'hour'
user_id?: number
api_key_id?: number
model?: string
account_id?: number
group_id?: number
stream?: boolean
}
export interface TrendResponse {
......@@ -70,6 +74,10 @@ export interface ModelStatsParams {
end_date?: string
user_id?: number
api_key_id?: number
model?: string
account_id?: number
group_id?: number
stream?: boolean
}
export interface ModelStatsResponse {
......
frontend/src/api/admin/ops.ts
View file @ 90bce60b
......@@ -17,6 +17,47 @@ export interface OpsRequestOptions {
export interface OpsRetryRequest {
mode: OpsRetryMode
pinned_account_id?: number
force?: boolean
}
export interface OpsRetryAttempt {
id: number
created_at: string
requested_by_user_id: number
source_error_id: number
mode: string
pinned_account_id?: number | null
pinned_account_name?: string
status: string
started_at?: string | null
finished_at?: string | null
duration_ms?: number | null
success?: boolean | null
http_status_code?: number | null
upstream_request_id?: string | null
used_account_id?: number | null
used_account_name?: string
response_preview?: string | null
response_truncated?: boolean | null
result_request_id?: string | null
result_error_id?: number | null
error_message?: string | null
}
export type OpsUpstreamErrorEvent = {
at_unix_ms?: number
platform?: string
account_id?: number
account_name?: string
upstream_status_code?: number
upstream_request_id?: string
upstream_request_body?: string
kind?: string
message?: string
detail?: string
}
export interface OpsRetryResult {
......@@ -626,8 +667,6 @@ export type MetricType =
| 'success_rate'
| 'error_rate'
| 'upstream_error_rate'
| 'p95_latency_ms'
| 'p99_latency_ms'
| 'cpu_usage_percent'
| 'memory_usage_percent'
| 'concurrency_queue_depth'
......@@ -663,7 +702,7 @@ export interface AlertEvent {
id: number
rule_id: number
severity: OpsSeverity | string
status: 'firing' | 'resolved' | string
status: 'firing' | 'resolved' | 'manual_resolved' | string
title?: string
description?: string
metric_value?: number
......@@ -702,7 +741,6 @@ export interface EmailNotificationConfig {
export interface OpsMetricThresholds {
sla_percent_min?: number | null // SLA低于此值变红
latency_p99_ms_max?: number | null // 延迟P99高于此值变红
ttft_p99_ms_max?: number | null // TTFT P99高于此值变红
request_error_rate_percent_max?: number | null // 请求错误率高于此值变红
upstream_error_rate_percent_max?: number | null // 上游错误率高于此值变红
......@@ -735,6 +773,8 @@ export interface OpsAdvancedSettings {
data_retention: OpsDataRetentionSettings
aggregation: OpsAggregationSettings
ignore_count_tokens_errors: boolean
ignore_context_canceled: boolean
ignore_no_available_accounts: boolean
auto_refresh_enabled: boolean
auto_refresh_interval_seconds: number
}
......@@ -754,21 +794,37 @@ export interface OpsAggregationSettings {
export interface OpsErrorLog {
id: number
created_at: string
// Standardized classification
phase: OpsPhase
type: string
error_owner: 'client' | 'provider' | 'platform' | string
error_source: 'client_request' | 'upstream_http' | 'gateway' | string
severity: OpsSeverity
status_code: number
platform: string
model: string
latency_ms?: number | null
is_retryable: boolean
retry_count: number
resolved: boolean
resolved_at?: string | null
resolved_by_user_id?: number | null
resolved_retry_id?: number | null
client_request_id: string
request_id: string
message: string
user_id?: number | null
user_email: string
api_key_id?: number | null
account_id?: number | null
account_name: string
group_id?: number | null
group_name: string
client_ip?: string | null
request_path?: string
......@@ -890,7 +946,9 @@ export async function getErrorDistribution(
return data
}
export async function listErrorLogs(params: {
export type OpsErrorListView = 'errors' | 'excluded' | 'all'
export type OpsErrorListQueryParams = {
page?: number
page_size?: number
time_range?: string
......@@ -899,10 +957,20 @@ export async function listErrorLogs(params: {
platform?: string
group_id?: number | null
account_id?: number | null
phase?: string
error_owner?: string
error_source?: string
resolved?: string
view?: OpsErrorListView
q?: string
status_codes?: string
}): Promise<OpsErrorLogsResponse> {
status_codes_other?: string
}
// Legacy unified endpoints
export async function listErrorLogs(params: OpsErrorListQueryParams): Promise<OpsErrorLogsResponse> {
const { data } = await apiClient.get<OpsErrorLogsResponse>('/admin/ops/errors', { params })
return data
}
......@@ -917,6 +985,70 @@ export async function retryErrorRequest(id: number, req: OpsRetryRequest): Promi
return data
}
export async function listRetryAttempts(errorId: number, limit = 50): Promise<OpsRetryAttempt[]> {
const { data } = await apiClient.get<OpsRetryAttempt[]>(`/admin/ops/errors/${errorId}/retries`, { params: { limit } })
return data
}
export async function updateErrorResolved(errorId: number, resolved: boolean): Promise<void> {
await apiClient.put(`/admin/ops/errors/${errorId}/resolve`, { resolved })
}
// New split endpoints
export async function listRequestErrors(params: OpsErrorListQueryParams): Promise<OpsErrorLogsResponse> {
const { data } = await apiClient.get<OpsErrorLogsResponse>('/admin/ops/request-errors', { params })
return data
}
export async function listUpstreamErrors(params: OpsErrorListQueryParams): Promise<OpsErrorLogsResponse> {
const { data } = await apiClient.get<OpsErrorLogsResponse>('/admin/ops/upstream-errors', { params })
return data
}
export async function getRequestErrorDetail(id: number): Promise<OpsErrorDetail> {
const { data } = await apiClient.get<OpsErrorDetail>(`/admin/ops/request-errors/${id}`)
return data
}
export async function getUpstreamErrorDetail(id: number): Promise<OpsErrorDetail> {
const { data } = await apiClient.get<OpsErrorDetail>(`/admin/ops/upstream-errors/${id}`)
return data
}
export async function retryRequestErrorClient(id: number): Promise<OpsRetryResult> {
const { data } = await apiClient.post<OpsRetryResult>(`/admin/ops/request-errors/${id}/retry-client`, {})
return data
}
export async function retryRequestErrorUpstreamEvent(id: number, idx: number): Promise<OpsRetryResult> {
const { data } = await apiClient.post<OpsRetryResult>(`/admin/ops/request-errors/${id}/upstream-errors/${idx}/retry`, {})
return data
}
export async function retryUpstreamError(id: number): Promise<OpsRetryResult> {
const { data } = await apiClient.post<OpsRetryResult>(`/admin/ops/upstream-errors/${id}/retry`, {})
return data
}
export async function updateRequestErrorResolved(errorId: number, resolved: boolean): Promise<void> {
await apiClient.put(`/admin/ops/request-errors/${errorId}/resolve`, { resolved })
}
export async function updateUpstreamErrorResolved(errorId: number, resolved: boolean): Promise<void> {
await apiClient.put(`/admin/ops/upstream-errors/${errorId}/resolve`, { resolved })
}
export async function listRequestErrorUpstreamErrors(
id: number,
params: OpsErrorListQueryParams = {},
options: { include_detail?: boolean } = {}
): Promise<PaginatedResponse<OpsErrorDetail>> {
const query: Record<string, any> = { ...params }
if (options.include_detail) query.include_detail = '1'
const { data } = await apiClient.get<PaginatedResponse<OpsErrorDetail>>(`/admin/ops/request-errors/${id}/upstream-errors`, { params: query })
return data
}
export async function listRequestDetails(params: OpsRequestDetailsParams): Promise<OpsRequestDetailsResponse> {
const { data } = await apiClient.get<OpsRequestDetailsResponse>('/admin/ops/requests', { params })
return data
......@@ -942,11 +1074,45 @@ export async function deleteAlertRule(id: number): Promise<void> {
await apiClient.delete(`/admin/ops/alert-rules/${id}`)
}
export async function listAlertEvents(limit = 100): Promise<AlertEvent[]> {
const { data } = await apiClient.get<AlertEvent[]>('/admin/ops/alert-events', { params: { limit } })
export interface AlertEventsQuery {
limit?: number
status?: string
severity?: string
email_sent?: boolean
time_range?: string
start_time?: string
end_time?: string
before_fired_at?: string
before_id?: number
platform?: string
group_id?: number
}
export async function listAlertEvents(params: AlertEventsQuery = {}): Promise<AlertEvent[]> {
const { data } = await apiClient.get<AlertEvent[]>('/admin/ops/alert-events', { params })
return data
}
export async function getAlertEvent(id: number): Promise<AlertEvent> {
const { data } = await apiClient.get<AlertEvent>(`/admin/ops/alert-events/${id}`)
return data
}
export async function updateAlertEventStatus(id: number, status: 'resolved' | 'manual_resolved'): Promise<void> {
await apiClient.put(`/admin/ops/alert-events/${id}/status`, { status })
}
export async function createAlertSilence(payload: {
rule_id: number
platform: string
group_id?: number | null
region?: string | null
until: string
reason?: string
}): Promise<void> {
await apiClient.post('/admin/ops/alert-silences', payload)
}
// Email notification config
export async function getEmailNotificationConfig(): Promise<EmailNotificationConfig> {
const { data } = await apiClient.get<EmailNotificationConfig>('/admin/ops/email-notification/config')
......@@ -1001,15 +1167,35 @@ export const opsAPI = {
getAccountAvailabilityStats,
getRealtimeTrafficSummary,
subscribeQPS,
// Legacy unified endpoints
listErrorLogs,
getErrorLogDetail,
retryErrorRequest,
listRetryAttempts,
updateErrorResolved,
// New split endpoints
listRequestErrors,
listUpstreamErrors,
getRequestErrorDetail,
getUpstreamErrorDetail,
retryRequestErrorClient,
retryRequestErrorUpstreamEvent,
retryUpstreamError,
updateRequestErrorResolved,
updateUpstreamErrorResolved,
listRequestErrorUpstreamErrors,
listRequestDetails,
listAlertRules,
createAlertRule,
updateAlertRule,
deleteAlertRule,
listAlertEvents,
getAlertEvent,
updateAlertEventStatus,
createAlertSilence,
getEmailNotificationConfig,
updateEmailNotificationConfig,
getAlertRuntimeSettings,
......
frontend/src/api/admin/proxies.ts
View file @ 90bce60b
......@@ -4,7 +4,13 @@
*/
import { apiClient } from '../client'
import type { Proxy, CreateProxyRequest, UpdateProxyRequest, PaginatedResponse } from '@/types'
import type {
Proxy,
ProxyAccountSummary,
CreateProxyRequest,
UpdateProxyRequest,
PaginatedResponse
} from '@/types'
/**
* List all proxies with pagination
......@@ -160,8 +166,8 @@ export async function getStats(id: number): Promise<{
* @param id - Proxy ID
* @returns List of accounts using the proxy
*/
export async function getProxyAccounts(id: number): Promise<PaginatedResponse<any>> {
const { data } = await apiClient.get<PaginatedResponse<any>>(`/admin/proxies/${id}/accounts`)
export async function getProxyAccounts(id: number): Promise<ProxyAccountSummary[]> {
const { data } = await apiClient.get<ProxyAccountSummary[]>(`/admin/proxies/${id}/accounts`)
return data
}
......@@ -189,6 +195,17 @@ export async function batchCreate(
return data
}
export async function batchDelete(ids: number[]): Promise<{
deleted_ids: number[]
skipped: Array<{ id: number; reason: string }>
}> {
const { data } = await apiClient.post<{
deleted_ids: number[]
skipped: Array<{ id: number; reason: string }>
}>('/admin/proxies/batch-delete', { ids })
return data
}
export const proxiesAPI = {
list,
getAll,
......@@ -201,7 +218,8 @@ export const proxiesAPI = {
testProxy,
getStats,
getProxyAccounts,
batchCreate
batchCreate,
batchDelete
}
export default proxiesAPI
frontend/src/api/admin/usage.ts
View file @ 90bce60b
......@@ -16,6 +16,7 @@ export interface AdminUsageStatsResponse {
total_tokens: number
total_cost: number
total_actual_cost: number
total_account_cost?: number
average_duration_ms: number
}
......
frontend/src/components/account/AccountStatsModal.vue
View file @ 90bce60b
......@@ -73,11 +73,12 @@
</p>
<p class="mt-1 text-xs text-gray-500 dark:text-gray-400">
{{ t('admin.accounts.stats.accumulatedCost') }}
<span class="text-gray-400 dark:text-gray-500"
>({{ t('admin.accounts.stats.standardCost') }}: ${{
<span class="text-gray-400 dark:text-gray-500">
({{ t('usage.userBilled') }}: ${{ formatCost(stats.summary.total_user_cost) }} ·
{{ t('admin.accounts.stats.standardCost') }}: ${{
formatCost(stats.summary.total_standard_cost)
}})</span
>
}})
</span>
</p>
</div>
......@@ -127,6 +128,9 @@
days: stats.summary.actual_days_used
})
}}
<span class="text-gray-400 dark:text-gray-500">
({{ t('usage.userBilled') }}: ${{ formatCost(stats.summary.avg_daily_user_cost) }})
</span>
</p>
</div>
......@@ -189,13 +193,17 @@
</div>
<div class="space-y-2">
<div class="flex items-center justify-between">
<span class="text-xs text-gray-500 dark:text-gray-400">{{
t('admin.accounts.stats.cost')
}}</span>
<span class="text-xs text-gray-500 dark:text-gray-400">{{ t('usage.accountBilled') }}</span>
<span class="text-sm font-semibold text-gray-900 dark:text-white"
>${{ formatCost(stats.summary.today?.cost || 0) }}</span
>
</div>
<div class="flex items-center justify-between">
<span class="text-xs text-gray-500 dark:text-gray-400">{{ t('usage.userBilled') }}</span>
<span class="text-sm font-semibold text-gray-900 dark:text-white"
>${{ formatCost(stats.summary.today?.user_cost || 0) }}</span
>
</div>
<div class="flex items-center justify-between">
<span class="text-xs text-gray-500 dark:text-gray-400">{{
t('admin.accounts.stats.requests')
......@@ -240,13 +248,17 @@
}}</span>
</div>
<div class="flex items-center justify-between">
<span class="text-xs text-gray-500 dark:text-gray-400">{{
t('admin.accounts.stats.cost')
}}</span>
<span class="text-xs text-gray-500 dark:text-gray-400">{{ t('usage.accountBilled') }}</span>
<span class="text-sm font-semibold text-orange-600 dark:text-orange-400"
>${{ formatCost(stats.summary.highest_cost_day?.cost || 0) }}</span
>
</div>
<div class="flex items-center justify-between">
<span class="text-xs text-gray-500 dark:text-gray-400">{{ t('usage.userBilled') }}</span>
<span class="text-sm font-semibold text-gray-900 dark:text-white"
>${{ formatCost(stats.summary.highest_cost_day?.user_cost || 0) }}</span
>
</div>
<div class="flex items-center justify-between">
<span class="text-xs text-gray-500 dark:text-gray-400">{{
t('admin.accounts.stats.requests')
......@@ -291,13 +303,17 @@
}}</span>
</div>
<div class="flex items-center justify-between">
<span class="text-xs text-gray-500 dark:text-gray-400">{{
t('admin.accounts.stats.cost')
}}</span>
<span class="text-xs text-gray-500 dark:text-gray-400">{{ t('usage.accountBilled') }}</span>
<span class="text-sm font-semibold text-gray-900 dark:text-white"
>${{ formatCost(stats.summary.highest_request_day?.cost || 0) }}</span
>
</div>
<div class="flex items-center justify-between">
<span class="text-xs text-gray-500 dark:text-gray-400">{{ t('usage.userBilled') }}</span>
<span class="text-sm font-semibold text-gray-900 dark:text-white"
>${{ formatCost(stats.summary.highest_request_day?.user_cost || 0) }}</span
>
</div>
</div>
</div>
</div>
......@@ -397,13 +413,17 @@
}}</span>
</div>
<div class="flex items-center justify-between">
<span class="text-xs text-gray-500 dark:text-gray-400">{{
t('admin.accounts.stats.todayCost')
}}</span>
<span class="text-xs text-gray-500 dark:text-gray-400">{{ t('usage.accountBilled') }}</span>
<span class="text-sm font-semibold text-gray-900 dark:text-white"
>${{ formatCost(stats.summary.today?.cost || 0) }}</span
>
</div>
<div class="flex items-center justify-between">
<span class="text-xs text-gray-500 dark:text-gray-400">{{ t('usage.userBilled') }}</span>
<span class="text-sm font-semibold text-gray-900 dark:text-white"
>${{ formatCost(stats.summary.today?.user_cost || 0) }}</span
>
</div>
</div>
</div>
</div>
......@@ -517,14 +537,24 @@ const trendChartData = computed(() => {
labels: stats.value.history.map((h) => h.label),
datasets: [
{
label: t('admin.accounts.stats.cost') + ' (USD)',
data: stats.value.history.map((h) => h.cost),
label: t('usage.accountBilled') + ' (USD)',
data: stats.value.history.map((h) => h.actual_cost),
borderColor: '#3b82f6',
backgroundColor: 'rgba(59, 130, 246, 0.1)',
fill: true,
tension: 0.3,
yAxisID: 'y'
},
{
label: t('usage.userBilled') + ' (USD)',
data: stats.value.history.map((h) => h.user_cost),
borderColor: '#10b981',
backgroundColor: 'rgba(16, 185, 129, 0.08)',
fill: false,
tension: 0.3,
borderDash: [5, 5],
yAxisID: 'y'
},
{
label: t('admin.accounts.stats.requests'),
data: stats.value.history.map((h) => h.requests),
......@@ -602,7 +632,7 @@ const lineChartOptions = computed(() => ({
},
title: {
display: true,
text: t('admin.accounts.stats.cost') + ' (USD)',
text: t('usage.accountBilled') + ' (USD)',
color: '#3b82f6',
font: {
size: 11
......
frontend/src/components/account/AccountTodayStatsCell.vue
View file @ 90bce60b
......@@ -32,15 +32,20 @@
formatTokens(stats.tokens)
}}</span>
</div>
<!-- Cost -->
<!-- Cost (Account) -->
<div class="flex items-center gap-1">
<span class="text-gray-500 dark:text-gray-400"
>{{ t('admin.accounts.stats.cost') }}:</span
>
<span class="text-gray-500 dark:text-gray-400">{{ t('usage.accountBilled') }}:</span>
<span class="font-medium text-emerald-600 dark:text-emerald-400">{{
formatCurrency(stats.cost)
}}</span>
</div>
<!-- Cost (User/API Key) -->
<div v-if="stats.user_cost != null" class="flex items-center gap-1">
<span class="text-gray-500 dark:text-gray-400">{{ t('usage.userBilled') }}:</span>
<span class="font-medium text-gray-700 dark:text-gray-300">{{
formatCurrency(stats.user_cost)
}}</span>
</div>
</div>
<!-- No data -->
......
frontend/src/components/account/BulkEditAccountModal.vue
View file @ 90bce60b
......@@ -459,7 +459,7 @@
</div>
<!-- Concurrency & Priority -->
<div class="grid grid-cols-2 gap-4 border-t border-gray-200 pt-4 dark:border-dark-600">
<div class="grid grid-cols-2 gap-4 border-t border-gray-200 pt-4 dark:border-dark-600 lg:grid-cols-3">
<div>
<div class="mb-3 flex items-center justify-between">
<label
......@@ -516,6 +516,36 @@
aria-labelledby="bulk-edit-priority-label"
/>
</div>
<div>
<div class="mb-3 flex items-center justify-between">
<label
id="bulk-edit-rate-multiplier-label"
class="input-label mb-0"
for="bulk-edit-rate-multiplier-enabled"
>
{{ t('admin.accounts.billingRateMultiplier') }}
</label>
<input
v-model="enableRateMultiplier"
id="bulk-edit-rate-multiplier-enabled"
type="checkbox"
aria-controls="bulk-edit-rate-multiplier"
class="rounded border-gray-300 text-primary-600 focus:ring-primary-500"
/>
</div>
<input
v-model.number="rateMultiplier"
id="bulk-edit-rate-multiplier"
type="number"
min="0"
step="0.01"
:disabled="!enableRateMultiplier"
class="input"
:class="!enableRateMultiplier && 'cursor-not-allowed opacity-50'"
aria-labelledby="bulk-edit-rate-multiplier-label"
/>
<p class="input-hint">{{ t('admin.accounts.billingRateMultiplierHint') }}</p>
</div>
</div>
<!-- Status -->
......@@ -655,6 +685,7 @@ const enableInterceptWarmup = ref(false)
const enableProxy = ref(false)
const enableConcurrency = ref(false)
const enablePriority = ref(false)
const enableRateMultiplier = ref(false)
const enableStatus = ref(false)
const enableGroups = ref(false)
......@@ -670,6 +701,7 @@ const interceptWarmupRequests = ref(false)
const proxyId = ref<number | null>(null)
const concurrency = ref(1)
const priority = ref(1)
const rateMultiplier = ref(1)
const status = ref<'active' | 'inactive'>('active')
const groupIds = ref<number[]>([])
......@@ -863,6 +895,10 @@ const buildUpdatePayload = (): Record<string, unknown> | null => {
updates.priority = priority.value
}
if (enableRateMultiplier.value) {
updates.rate_multiplier = rateMultiplier.value
}
if (enableStatus.value) {
updates.status = status.value
}
......@@ -923,6 +959,7 @@ const handleSubmit = async () => {
enableProxy.value ||
enableConcurrency.value ||
enablePriority.value ||
enableRateMultiplier.value ||
enableStatus.value ||
enableGroups.value
......@@ -977,6 +1014,7 @@ watch(
enableProxy.value = false
enableConcurrency.value = false
enablePriority.value = false
enableRateMultiplier.value = false
enableStatus.value = false
enableGroups.value = false
......@@ -991,6 +1029,7 @@ watch(
proxyId.value = null
concurrency.value = 1
priority.value = 1
rateMultiplier.value = 1
status.value = 'active'
groupIds.value = []
}
......
frontend/src/components/account/CreateAccountModal.vue
View file @ 90bce60b
......@@ -1196,7 +1196,7 @@
<ProxySelector v-model="form.proxy_id" :proxies="proxies" />
</div>
<div class="grid grid-cols-2 gap-4">
<div class="grid grid-cols-2 gap-4 lg:grid-cols-3">
<div>
<label class="input-label">{{ t('admin.accounts.concurrency') }}</label>
<input v-model.number="form.concurrency" type="number" min="1" class="input" />
......@@ -1212,6 +1212,11 @@
/>
<p class="input-hint">{{ t('admin.accounts.priorityHint') }}</p>
</div>
<div>
<label class="input-label">{{ t('admin.accounts.billingRateMultiplier') }}</label>
<input v-model.number="form.rate_multiplier" type="number" min="0" step="0.01" class="input" />
<p class="input-hint">{{ t('admin.accounts.billingRateMultiplierHint') }}</p>
</div>
</div>
<div class="border-t border-gray-200 pt-4 dark:border-dark-600">
<label class="input-label">{{ t('admin.accounts.expiresAt') }}</label>
......@@ -1832,6 +1837,7 @@ const form = reactive({
proxy_id: null as number | null,
concurrency: 10,
priority: 1,
rate_multiplier: 1,
group_ids: [] as number[],
expires_at: null as number | null
})
......@@ -2119,6 +2125,7 @@ const resetForm = () => {
form.proxy_id = null
form.concurrency = 10
form.priority = 1
form.rate_multiplier = 1
form.group_ids = []
form.expires_at = null
accountCategory.value = 'oauth-based'
......@@ -2272,6 +2279,7 @@ const createAccountAndFinish = async (
proxy_id: form.proxy_id,
concurrency: form.concurrency,
priority: form.priority,
rate_multiplier: form.rate_multiplier,
group_ids: form.group_ids,
expires_at: form.expires_at,
auto_pause_on_expired: autoPauseOnExpired.value
......@@ -2490,6 +2498,7 @@ const handleCookieAuth = async (sessionKey: string) => {
proxy_id: form.proxy_id,
concurrency: form.concurrency,
priority: form.priority,
rate_multiplier: form.rate_multiplier,
group_ids: form.group_ids,
expires_at: form.expires_at,
auto_pause_on_expired: autoPauseOnExpired.value
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment