- Separate clientAffinityTTL (24h) from stickySessionTTL (1h)
- Atomic Lua dual-write for forward (client→accounts) and reverse
(account→clients) affinity indexes with lazy expiry cleanup
- Extract Lua scripts to standalone .lua files with //go:embed
- Add filterByMinAffinityCount to Layer 1 & 2 scheduling so new
clients are routed to accounts with fewest existing affinities
- Skip Redis query when no affinity-enabled accounts exist (zero
overhead for non-affinity users)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Split compound "Profile" column into separate Profile ID and Name columns
- Replace "Endpoint" column with "Storage Path" showing bucket/prefix or folder info
- Add GDrive quota display (capacity/used) via new backend API endpoint
- Add video count statistics per storage type via new backend API endpoint
- Add inline test button per profile with 15s timeout and per-profile loading state
- Backend: GetQuotaInfo, CountByStorageType, GetGDriveQuota, GetStorageVideoStats
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Replace single-period-selector (三选一) with three independently configurable
quota dimensions (日/周/总) that can all be set simultaneously. Account is
paused when ANY dimension is exceeded. Periodic quotas (daily/weekly) use
lazy reset via atomic SQL — no cron job needed.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
CTE snapshot in READ COMMITTED can cause concurrent requests during
period reset to lose one increment. Inline CASE in single UPDATE
ensures PostgreSQL re-evaluates all expressions with latest row data.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Use NOW() AT TIME ZONE 'UTC' to ensure consistent timestamps
regardless of database timezone setting
- Clear quota_period_start when manually resetting quota so a
fresh period starts from next usage
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Support daily/weekly automatic quota reset using a rolling period
anchored to first usage. Implements lazy reset via atomic CTE SQL
in IncrementQuotaUsed - no cron job needed.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Remove frontend badge/tooltip and backend counting/query infrastructure
for client affinity. Core client affinity scheduling logic is preserved.
Removed:
- GatewayCache injection in AccountHandler
- GetAccountAffinityCount/Batch/Clients methods and interface
- AffinityClientInfo struct and ClientAffinityTTL function
- /affinity-clients API endpoint and route
- Reverse index (account_affinity:*) Lua scripts
- Frontend blue badge, tooltip, and lazy-load logic
- affinity_client_count field from types and response
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The reverse index account_affinity:{accountID} was using the same 1h
EXPIRE as the forward index. But the forward key EXPIRE refreshes on
every client request, while the reverse key only refreshes when a
request is routed to that specific account. This caused reverse keys
to expire prematurely.
Fix: set reverse key EXPIRE to 24h as safety net, rely on
ZREMRANGEBYSCORE for member-level expiration. Also clean up empty
keys after member cleanup.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Backend: remove count > 0 filter in GetAccountAffinityCountBatch so
accounts with 0 bound clients are included in the response
- Frontend: change badge condition from affinityClientCount > 0 to
account.client_affinity_enabled so the badge always shows for accounts
with affinity enabled
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add reverse Redis index (account_affinity:{accountID}) to track which
clients are bound to each account via affinity scheduling. Display a
blue badge with client count in the admin account list, with a hover
tooltip showing client details (groupID:clientID + TTL).
Backend: reverse index in Lua scripts, batch query methods, new
GET /accounts/:id/affinity-clients endpoint.
Frontend: blue badge + lazy-loaded tooltip in AccountStatusIndicator.
- Separate load factor from concurrency: concurrency controls actual
slot acquisition, load_factor controls load rate calculation
- Add EffectiveLoadFactor() method: LoadFactor > Concurrency > 1
- Add load_factor field to Create/Edit/BulkEdit account forms
- Fix RPM default value: auto-fill 15 when RPM enabled but not set
- Fix stale test compilation errors in server and handler packages
Upstream errors like 401/429 must not be passed through as HTTP status
codes because the frontend routes on status (401 triggers JWT logout).
Keep HTTP 500 but include upstream error details in the message.
- Frontend: queue Anthropic OAuth/setup-token usage requests by proxy
with random 1-1.5s interval to prevent upstream 429
- Backend: return ApplicationError with actual upstream status code
instead of wrapping all errors as 500
- Handle component unmount to skip stale updates on page navigation
- Fix fileChecksum for 061 migration: use TrimSpace hash (66207e7a) instead
of raw sha256sum (97bdd9a3), matching the actual runtime computation
- Add 222b4a09 as accepted DB checksum for 061 migration
- Add missing GetAPIKeyRateLimit/SetAPIKeyRateLimit/UpdateAPIKeyRateLimitUsage/
InvalidateAPIKeyRateLimit methods to mock BillingCache in test stubs
- Fix NewBillingCacheService call in singleflight test (add apiKeyRepo param)
Add a doughnut chart showing usage statistics broken down by group on
the admin usage records page. The chart appears alongside the existing
model distribution chart (2-column grid), with the token usage trend
chart moved to a separate full-width row below.
Changes:
- backend/pkg/usagestats: add GroupStat type
- backend/service: add GetGroupStatsWithFilters interface method and implementation
- backend/repository: implement GetGroupStatsWithFilters with LEFT JOIN groups
- backend/handler: add GetGroupStats handler with full filter support
- backend/routes: register GET /admin/dashboard/groups route
- backend/tests: add GetGroupStatsWithFilters stubs to contract/sora tests
- frontend/types: add GroupStat interface
- frontend/api: add getGroupStats API function and types
- frontend/components: add GroupDistributionChart.vue doughnut chart
- frontend/views: update UsageView layout and load group stats in parallel
- frontend/i18n: add groupDistribution, group, noGroup keys (zh + en)
- Add GroupStat type to usagestats package
- Add GetGroupStatsWithFilters to UsageLogRepository interface and implement with LEFT JOIN groups
- Add GetGroupStats dashboard API endpoint (GET /admin/dashboard/groups)
- Add GroupDistributionChart.vue component mirroring ModelDistributionChart
- Rearrange UsageView layout: model + group in one row, token trend full-width below
- All filters (user, api_key, account, group, model, date range) apply to group stats
- Use TxPipeline (MULTI/EXEC) instead of Pipeline for atomic INCR+EXPIRE
- Filter negative values in GetBaseRPM(), update test expectation
- Add RPM batch query (GetRPMBatch) to account List API
- Add warn logs for RPM increment failures in gateway handler
- Reset enableRpmLimit on BulkEditAccountModal close
- Use union type 'tiered' | 'sticky_exempt' for rpmStrategy refs
- Add design decision comments for rdb.Time() RTT trade-off