Vexell

System status

All systems operational.

Live overview of every Vexell service — uptime windows, recent incidents, scheduled maintenance. Subscribe to get incident updates by email or RSS.

All services runningLast checked 2m ago · auto-refresh every 30sSubscribeRSS

API

api.vexell.com · global

Operational

Dashboard

app.vexell.com

Operational

AI engines

GPT-4o, Claude Sonnet 4.6

Degraded performance

Integrations

GitHub, Slack, Linear

Operational

Uptime · last 90 days

One cell per day. Hover for the daily summary.

API

REST + streaming endpoints (api.vexell.com)

99.94% uptime
90 days agoLast 90 daysToday

Dashboard

app.vexell.com — login, runs, settings

99.97% uptime
90 days agoLast 90 daysToday

AI engines

Vendor-passthrough (GPT-4o, Claude Sonnet 4.6, Llama 3.1)

99.78% uptime
90 days agoLast 90 daysToday

Integrations

GitHub, Slack, Linear, Sentry

100% uptime
90 days agoLast 90 daysToday

Recent incidents

Latest at the top. Click any incident to expand the investigation timeline.

See archive

Elevated latency on AI engines

May 1, 2026 · 09:14 UTC

Affected: AI engines · GPT-4o

Resolved · 38 min

  1. resolved

    GPT-4o vendor recovered. p99 back below 1.4s. We rebalanced traffic across providers and added a circuit breaker for next time.

    May 1, 09:52 UTC

  2. identified

    Vendor (OpenAI) reports degraded performance on GPT-4o in us-east. Routing affected requests to Claude Sonnet fallback where appropriate.

    May 1, 09:31 UTC

  3. investigating

    p99 latency on /v1/runs spiked from 1.2s → 4.8s in the last 5 minutes. Investigating upstream model providers.

    May 1, 09:14 UTC

GPT-4o vendor outage

Apr 3, 2026 · 14:08 UTC

Affected: AI engines · GPT-4o

Resolved · 1h 02m

  1. resolved

    OpenAI confirmed full recovery. Backfilled queued requests. No data loss.

    Apr 3, 15:10 UTC

  2. update

    Auto-fallback to Claude Sonnet engaged for 87% of incoming runs. The remaining 13% (vision-required) are queued.

    Apr 3, 14:32 UTC

  3. investigating

    GPT-4o requests returning 5xx across all regions. OpenAI status confirms major outage.

    Apr 3, 14:08 UTC

Slow dashboard loads

Apr 22, 2026 · 11:47 UTC

Affected: Dashboard

Resolved · 21 min

  1. resolved

    Cleared. Root cause: a stuck connection to Postgres replicas; we cycled the pool and added a tighter health check.

    Apr 22, 12:08 UTC

  2. monitoring

    Initial loads of /dashboard taking 4-7s instead of <1s. Investigating downstream cache.

    Apr 22, 11:47 UTC

Get incident updates

Email when we open a new incident, when status changes, and when it’s resolved. We don’t use this for anything else.