Monitor Telegraf Controller

Monitor a Telegraf Controller server with its built-in health endpoints and its log output.

• Health endpoints
• Server logs
• What to watch for

Health endpoints

Telegraf Controller exposes unauthenticated HTTP health endpoints on the API port and on the heartbeat port. GET /health/live and GET /health/ready on the API port are the general-purpose liveness and readiness probes, and the heartbeat service answers GET /health on its own port. GET /health/leader identifies the leader in a high-availability cluster.

For the full endpoint reference, including status codes, response bodies, and how to choose a probe, see Health endpoints. The endpoints are documented with load balancers in mind, but any monitoring system can poll them.

Server logs

Telegraf Controller writes log output to the console (standard output and standard error). Where that output ends up depends on how you run the server:

• Interactive terminal: log output appears in the terminal.
• systemd: the journal captures console output. Read it with journalctl -u telegraf-controller.
• LaunchDaemon (macOS): output goes to the file paths configured in the service plist. The LaunchDaemon example uses /var/log/telegraf-controller.log.
• Windows service: configure your service manager to capture console output to files. See Run Telegraf Controller as a service.

What to watch for

Alert on the following log messages. Each entry links to the page that explains how to fix the problem.

• Database errors, such as database is locked, database or disk is full, or database disk image is malformed. See Troubleshoot the database.
• Rejected agent heartbeats, which return HTTP 401 with an invalid token error. See Agent heartbeats return 401 Invalid token.
• License errors. See Troubleshoot licensing.