Hello Prometheus community,
We are excited to announce the release of Prometheus 3.11.0-rc.0.
This release adds new PromQL operators for native histograms: `</` and `>/`
allow trimming observations from native histograms, and the new experimental
`histogram_quantiles` variadic function enables computing multiple quantiles at
once in a single expression.
On the storage side, a new `storage.tsdb.retention.percentage` configuration
option lets you cap TSDB disk usage to a percentage of the available disk space.
Several experimental TSDB feature flags are also introduced: `fast-startup` for
faster restarts via a series state cache, `st-storage` for persisting start
timestamps through Remote Write 2, and `xor2-encoding` for a new chunk encoding
optimised for scraped data.
Service discovery sees new AWS roles for Elasticache and RDS, Azure Workload
Identity authentication support, and Kubernetes SD improvements including node
role selectors for pod roles and new pod labels for deployment, cronjob, and job
controller names.
This release also includes multiple performance improvements across PromQL
(joins, native histogram aggregations, regex matchers) and Remote Write (WAL
watcher buffer reuse), along with many bug fixes.
You can find the full changelog and download Prometheus 3.11.0-rc.0 at
https://github.com/prometheus/prometheus/releases/tag/v3.11.0-rc.0.
Container images are also available at
https://quay.io/repository/prometheus/prometheus?tab=tags and
https://hub.docker.com/r/prom/prometheus/tags.
Thank you for your contributions and support in making Prometheus a
better tool for monitoring and alerting.
Best regards,
--
Julien Pivotto
@roidelapluie