justin
|
a4a5500bfc
|
Add Prometheus + Grafana + Alertmanager monitoring stack
Full observability stack with Telegram alerting:
Components:
- Prometheus: metrics collection, 90-day retention
- Grafana: dashboards at monitoring.performancewest.net
- Alertmanager: routes alerts to Telegram bot
- node-exporter: OS metrics (CPU, RAM, disk, network)
- cAdvisor: container metrics (CPU, memory, restarts)
- postgres-exporter: PostgreSQL connection/query metrics
- nginx-exporter: request rate, 5xx errors, connections
- blackbox-exporter: HTTP/TCP endpoint probing + SSL cert checks
Alert rules:
- Service down (HTTP probe, TCP port, container missing)
- Container restart loops
- High CPU/memory/disk/load
- PostgreSQL down or high connections
- SSL cert expiring (14d warning, 3d critical)
- Slow HTTP responses, high 5xx rate
Blackbox probes all public endpoints:
performancewest.net, api, dev, crm, lists, analytics,
minio, crypto, pay
Telegram alerts: critical=1h repeat, warning=6h repeat,
auto-resolve notifications
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
2026-05-01 02:08:39 -05:00 |
|