Telegraf v1.38.1 released on 16-03-2026
Telegraf v1.38.1 is out now. The release fixes a range of bugs that reduce data and label corruption, improve observability, and correct platform-specific metric collection.
For downloads, SHA256 checksums and the full changelog, consult the project’s GitHub release and the published release assets.
What’s in this release
- inputs.diskio: newline characters in the serial tag are sanitised to avoid invalid or fragile tag values.
- inputs.docker: status metrics are now emitted for non-running containers so stopped and paused containers report state.
- Operator-facing fixes and tidy-ups: inputs.exec now surfaces stderr to logs; inputs.mem uses vm.Cached as vm.Buffers on OpenBSD; inputs.ping warns when a timeout parameter is used with the native method; internal Go version extraction is more robust; outputs.influxdb_v3 had a duplicate timeout setting removed.
Upgrade notes
- Metric and tag changes can affect dashboards and alerts. The docker input may introduce new series, and disk serial sanitisation can change tag values—test upgrades in staging and update alert rules that assume container-count semantics or specific serial tags.
- If you vendor or embed Telegraf, rebuild and run integration tests because of dependency bumps. Downloads include SHA256 checksums—verify them before installing. To roll back, reinstall the previous package for your OS/architecture from your own archives or the prior release assets.
Share any feedback or issues on the GitHub release thread or your usual community channels — tell others what you saw after upgrading.
