Stop usage analytics reads from backfilling missing API key attribution
into historical rows so legacy records remain unchanged and surface as
unknown instead of guessed ownership.
Also require management authentication on admin concurrency and
compression analytics endpoints, and preserve image payloads during lite
compression unless a model is explicitly marked as non-vision.
Addresses chatgpt-codex-connector review on #1828: reconcileEnabledConnections
created auto-enabled limiters via direct \`new Bottleneck(...)\`, bypassing
the \`executing\` event listener that updates lastDispatchAt. The watchdog
then read \`lastDispatchAt.get(key) ?? 0\` and computed \`stalledMs = now - 0\`
≈ Unix time, force-resetting healthy idle limiters during normal reservoir-
refresh waits.
- Reconcile path now calls getLimiter(provider, connectionId), which registers
both queued/executing listeners and seeds lastDispatchAt to now.
- Watchdog also hardened: if lastDispatch is undefined for any reason, seed
it on the current tick and skip — never compute stall against epoch.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Addresses gemini-code-assist review on #1828: at 5s, the watchdog can
false-trigger on a healthy limiter that's correctly waiting on its
reservoir refresh (60s default) or adaptive minTime (up to ~60s for
1-RPM providers). 120s gives a 2× margin against both while still
catching the actual wedge case (observed at 3+ minutes stalled).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The auto-enabled Bottleneck safety net for API-key providers can desync
its internal state and silently stop dispatching jobs (queued > 0,
running == 0, executing == 0). Once that happens nothing recovers without
a process restart. This change adds a self-heal layer plus operator
visibility, and keeps existing behavior intact by default.
- Watchdog (30s tick) detects wedged limiters and force-resets them with
stop({dropWaitingJobs:true}) so queued callers actually fail rather than
stall forever.
- RATE_LIMIT_AUTO_ENABLE env var overrides the dashboard auto-enable
toggle (highest precedence). Lets operators flip the safety net off
during an incident without needing dashboard access.
- disableRateLimitProtection now uses stop({dropWaitingJobs:true})
instead of disconnect(); disconnect leaks queued promises (observed
while debugging this).
- STAGE_TRACE checkpoints in chatCore (post_injection, post_translation,
pre/post_semaphore, pre/inside/post_rate_limit, pre/post_executor)
pinpoint which await a hung request was stuck on.
- New /api/admin/concurrency endpoint exposes per-limiter counts and
semaphore stats so operators can see liveness in real time.
- Test coverage for the env-var override.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* feat(api-keys): add rename support in permissions modal
Add an editable key name field at the top of the permissions modal,
allowing users to rename API keys alongside existing permission settings.
The backend already supported name updates via PATCH /api/keys/:id — this
wires the UI to send the name field and refreshes the key list on success.
Changes:
- Add keyName state and text input to PermissionsModal
- Update handleUpdatePermissions to validate and send name in PATCH body
- Add integration test for rename via PATCH (valid, empty, too-long names)
- Update E2E mock to handle PATCH requests
* chore(release): bump version to 3.7.6
* chore(release): v3.7.6 — merge API key rename feature and sync docs
* chore(release): expand contributor credits to 155 PRs across full project history
- Expanded acknowledgment table from 29 to 53 contributors
- Added 100+ previously uncredited PRs from project inception through v3.7.5
- Moved contributor credits section to v3.7.6 (current release)
- Synced llm.txt version to 3.7.6
* fix: resolve security ReDoS in codex and bugs #1797#1789
* feat(dashboard): implement remaining v3.7.6 dashboard features and fixes
* fix(xiaomi-mimo): update models to V2.5, fix Token Plan validation and default region (#1823)
Integrated into release/v3.7.6
* fix(dashboard): correct loadPresets ReferenceError in CostOverviewTab
* fix(codex): omit compact client metadata (#1822)
Integrated into release/v3.7.6
* feat(chatgpt-web): support thinking_effort (Standard/Extended) for thinking-capable models (#1821)
Integrated into release/v3.7.6
* Fix endpoint visibility, A2A status, and API catalog (#1806)
Integrated into release/v3.7.6
* fix(analytics): use pure SQL aggregations — no history rows loaded (#1802)
Integrated into release/v3.7.6
* fix(stability): resolve codex input validation, enable combo circuit breaker, and fix broken unit tests
* docs(changelog): update for stability bug fixes#1804#1805
* fix: clear active requests and recover providers (#1824)
Integrated into release/v3.7.6
* feat: inject fallback tool names to prevent upstream 400 errors (#1775)
* feat: auto-restore probe-failed database to prevent data loss (#1810)
* fix: safely cast inputs to strings before calling trim() to avoid crashes on numeric fields in proxy modal (#1825)
* chore(release): v3.7.6 — final stability patches for production
* test: update expected db probe-failure error message for auto-restore feature
* chore(workflow): mandate implementation plan generation in resolve-issues
* docs(changelog): rewrite v3.7.6 with complete commit-accurate entries
* feat(analytics): add cost-based usage insights and activity streaks
Expand usage analytics to report total cost, per-series cost totals,
API key counts, and current activity streaks using pricing-aware token
calculations.
Also make probe-failed database recovery choose the newest backup by
its embedded timestamp instead of filesystem mtime so auto-restore
selects the intended snapshot reliably.
* fix(mitm): enforce transparent interception on port 443 only
Reject non-443 MITM port updates in the settings API and normalize
stored configuration back to the required transparent interception
port.
Lock the dashboard port field to 443, update the validation copy, and
add integration coverage to prevent stale custom ports from being
accepted or surfaced.
* docs(changelog): update for analytics and mitm features
---------
Co-authored-by: Andrew Munsell <andrew@wizardapps.net>
Co-authored-by: Antigravity Assistant <bot@antigravity.local>
Co-authored-by: Gi99lin <74502520+Gi99lin@users.noreply.github.com>
Co-authored-by: Sergey Morozov <tr0st@bk.ru>
Co-authored-by: payne <baboialex95@gmail.com>
Co-authored-by: Randi <55005611+rdself@users.noreply.github.com>
Co-authored-by: Paijo <14921983+oyi77@users.noreply.github.com>
Co-authored-by: ipanghu <bypanghu@163.com>
