Prometheus 2.18.0 is now out, following on from 2.17.0 with many fixes and improvements.

This is a fairly light release. The most notable change is that the amount of WAL kept on disk has been halved from 6 to 3 hours, which came out of my research for a previous blog post.

Federation now only queries the local TSDB, which should have already been the case for the majority of remote read users due to read_recent defaulting to false. EC2 service discovery now provides the architecture available as metadata.

There’s also a number of improvements, bug fixes and other changes, see the release notes for more detail.


Wondering how to upgrade? Contact us.