InfluxDB Monitoring, Backup, and Restore Guide

InfluxDB includes a built‑in monitoring system that, by default, collects system metrics every 10 seconds and stores them in the _internal database under the monitor retention policy (7‑day retention).

The _internal database contains several measurements such as cq , database , httpd , queryExecutor , runtime , shard , subscriber , tsm1_cache , tsm1_engine , tsm1_filestore , tsm1_wal , and write . An example query to view the latest write statistics is:

select * from "write" order by time desc limit 1

For detailed field definitions, refer to the official InfluxDB documentation.

Common InfluxDB commands include SHOW STATS (optionally filtered by a component, e.g., show stats for 'queryExecutor' ) and SHOW DIAGNOSTICS , which returns version, uptime, host, configuration, memory usage, and Go runtime information without storing the results in _internal .

System information can also be retrieved via HTTP endpoints /metrics (Go runtime metrics) and /debug/vars (similar to _internal data).

Backup and restore are performed over TCP. For remote backups, adjust the bind-address in the configuration to an externally reachable address and use the -host flag when running the commands.

Backup command example:

Restore command example:

Both commands share similar parameters; -portable selects the newer backup format (recommended) while -online uses the legacy format. Backups are full‑snapshot only; incremental backups are not supported. Time‑range options ( -start , -end ) operate on data blocks, which may include points outside the specified range and can overwrite existing points on restore.

Restoring directly into an existing database or retention policy is not allowed; a temporary database/RP must be used first, after which data can be moved into the target using a SELECT ... INTO statement.