Telegraf v1.40.0 released on 07-09-2026
Telegraf v1.40.0 is out now. Network telemetry users, SMART storage monitors and teams using parquet or removed plugins should review metric names, units and configuration.
Check the GitHub release notes and the official downloads page for full details, packages and SHA256 checksums before upgrading.
What’s in this release
- Cisco Telemetry MDT fixes: inputs.cisco_telemetry_mdt now respects configured aliases, avoids spurious RIB metrics, emits correct row numbers for table data and prevents tags leaking into parent metrics.
- SMART input unit fix and removals: inputs.smart now reports power_on_hours in hours for drives that expose the attribute as a duration (was previously seconds); several deprecated plugins and options were removed (notable examples: inputs.aerospike, inputs.sflow, outputs.amon, inputs.tail from_beginning and removed kafka_consumer/openstack/procstat options).
- New plugins and feature improvements: inputs.fritzbox_smarthome and outputs.zerobus added; secret store changes to remove secrets and support Vault namespaces; elasticsearch_query gained ES v7–v9 support; assorted enhancements such as configurable pod name tag for prometheus input, expanded nvidia_smi fields and additional proxmox disk/network I/O metrics.
Upgrade notes
- Breaking changes: several deprecated plugins/options have been removed (see examples above) and Cisco MDT changes can alter metric names or tags — audit your Telegraf config and update or remove affected blocks and dashboard queries before upgrading.
- If you need to revert, deploy the previous Telegraf package and verify package SHA256 checksums. Rebuild any container images that vendor dependencies if you maintain vendored builds.
Share any notes or issues you hit while upgrading — particularly around Cisco MDT metrics, SMART power_on_hours, or parquet behaviour — so others can learn from the experience.
