All systems operational
Real-time health of every VxCloud endpoint — frontend routes, backend APIs, the AI/LLM backend, and the Temporal workflow engine. Probed continuously from multiple regions.
Past 90 days
Incident history
Temporal workflow engine upgraded to 1.24. Workflows paused and resumed automatically; no task loss.
A model-cache warm-up after a node restart raised p95 inference latency. Mitigated by pre-warming on deploy.
Nginx upstream keep-alive exhaustion during a traffic spike. Tuned worker connections and upstream pool.
Continuous probes
Every endpoint is health-checked every few seconds from 5 regions, with TLS and upstream verification.
Auto-failover
Cloudflare routes to the healthy origin; the Go provisioner reschedules workloads on node loss.
Check from the CLI
Run vxcli status --watch to stream the same health you see here, straight into your terminal.