ha

High Availability (HA) Overview

• Run multiple proxy instances behind a load balancer; keep them stateless.
• Centralize rules via the API or Git (mounted config/rules.json) and reload on change.
• Prefer sticky connections only when TLS termination occurs at the proxy; otherwise TCP pass‑through is safe.
• Health probes: /healthz; readiness may include a quick upstream connect test.
• Metrics scraping: /metrics/prom for Prometheus; ship dashboards in docs/metrics-dashboard.md.

Notes

• Start with dry‑run in production; gradually enable enforcement by table/column.
• Use SECRET_PROVIDER=file with mounted secret files or cloud secret stores via sidecars.