Monitor health

The Node runtime exposes public health endpoints and an admin control API.

Public endpoints

GET /healthz

Returns JSON:

• ok
• service (synkd)
• version
• uptimeMs
• rooms
• connections
• metrics counters

Use this for load balancer health checks.

GET /metrics

Returns Prometheus text. Series names start with synkd_, for example synkd_info and synkd_connections.

Point your scraper at this path. Terminate TLS at the edge if the scrape is remote.

Admin control API

All /v1/* routes need an admin Bearer token.

Method	Path	Purpose
GET	/v1/health	Fleet health with tenant context
GET	/v1/rooms	Warm rooms
GET	/v1/capabilities	Runtime feature flags
GET	/v1/metrics	Same Prometheus body as /metrics
POST	/v1/rooms/:party/:room/hibernate	Force unload

Example:

curl -s -H "Authorization: Bearer dev-admin" \
  http://127.0.0.1:8787/v1/rooms

Logs

Set:

export SYNKD_LOG_LEVEL=info
export SYNKD_LOG_JSON=1

JSON logs are easier to ship to a log platform.

What to do next

• Wire scrapes after Deploy with TLS
• Review Security before opening ports