Metrics and Health

Health endpoints

PoolSwitch exposes:

• GET /health
• GET /status
• GET /metrics

/health is the lightweight readiness endpoint for local checks and load balancers.

Status endpoint

GET /status returns a JSON snapshot of the current key pool.

Typical fields:

• strategy
• storage backend
• upstream base URL
• key id
• total requests
• error count
• failover count
• estimated remaining quota
• last used time
• cooldown timestamp

Metrics endpoint

GET /metrics returns Prometheus-compatible metrics.

PoolSwitch currently tracks:

• poolswitch_requests_total
• poolswitch_failovers_total
• poolswitch_key_usage_total
• poolswitch_key_errors_total
• poolswitch_request_latency_seconds
• poolswitch_key_cooldown_active

Typical production setup

In production, teams usually:

• scrape /metrics with Prometheus
• alert when failovers spike
• alert when all keys are in cooldown
• monitor request latency and error rates

Embedded mode note

In embedded mode, the same metrics objects still exist inside the process, but there is no HTTP /metrics route unless you also run the proxy server.