Troubleshooting¶
Container is unhealthy or stops¶
Inspect the logs first (use the Compose file of your backend):
| Symptom | Cause | Action |
|---|---|---|
Running, unhealthy, log says permanent configuration error | The daemon exited with status 78; the watchdog is parked. The log line above names the offending option, or a configuration file that cannot be read; see Validation. | Fix .env, then docker compose restart fritzfluxdb. |
Log says rejected the configured credentials, not found, or TR-064 access is disabled | Startup preflight: the FRITZ!Box or database rejected the credentials, the InfluxDB database, organization or bucket is missing, or TR-064 is switched off. See Startup preflight. | Correct the setting (during operation, the FRITZ!Box rejections apply after 3 consecutive rejected login attempts, also when the password was changed and every TR-064 action answers 401; a 401 on only some actions, or on all of them while the login is accepted, means the user lacks the right for them and never stops the daemon), or enable Home Network > Network > Network Settings > Allow access for applications in the FRITZ!Box, then restart the container. |
Exits after repeated restarts, with the status of the last daemon failure (usually 1) | The daemon failed WATCHDOG_MAX_RESTARTS times in a row. | Read the last error in the log and fix its cause. |
No measurements arrive¶
Check the following in order:
- The FRITZ!Box host and credentials are correct and reachable from the container.
- The configured database host resolves from the container. In the bundled QuestDB example it is
questdb; the InfluxDB Compose examples require an existing endpoint and network connection. - The database, bucket, organization, or table permissions allow writes.
- TLS settings match the endpoint and certificate trust.
- The measurement/table name includes the sanitized FRITZ!Box serial when available.
Repeated backend warnings¶
The writer retries transient HTTP failures with backoff. Repeated warnings usually indicate reachability, TLS, authentication, or capacity problems. Persistent failures can eventually fill the bounded buffer and drop the oldest data. Retry rules are listed in InfluxDB.
Grafana shows no data¶
Confirm the datasource type and UID mapping, then use the actual serial-derived measurement name. The supplied dashboards do not invent data or query a fixed fritzbox name after a serial has been detected.
FRITZ!Box data is missing¶
Lua collection is disabled for FRITZ!OS versions below 7. Some services are model- or firmware- dependent. Look for service-specific warnings and compare the observed cadence with Polling and data model.