Prometheus 2.26.0 new features

Prometheus ​ ​2.26.0​​ After that, many repairs and improvements have been made .

Prometheus It is now built and supported Go 1.16 edition . This will resume Go 1.12 Memory release mode added in . This makes common RSS Use indicators to display more accurately Prometheus Actual memory used . Be careful , In this version ,Prometheus By default Alertmanager v2 edition . current Prometheus and Promtool Binary file printing help information and usage information will be output to stdout, instead of stderr , This result was agreed at the development summit . When using ​​--enable-feature=exemplar-storage​​ Parameter will store the sample data in memory later .

For the remote part , Yes ​​remote_writ​​ Added AWS SigV4 Support for authentication methods . Allow for ​​remote_write​​ Of HTTP 429 Response code to retry , But the retry action is disabled by default . In addition, it is allowed to be ​​remote_read​​ Configure customization header head . There are details of these changes , Please refer to the configuration documentation .

This time, a new version based on Scaleway Service discovery of ,Scaleway The slogan is “The Cloud that makes sense” . For the previous Digital Ocean Service discovery , I added a new one label ​​__meta_digitalocean_image​​.

about PromQL grammar , In the use of ​​--enable-feature=promql-negative-offset​​ Parameter allows the data to send a negative offset . In addition, this version adds ​​last_over_time​​, ​​sgn​​, ​​clamp​​ function , And the parser supports special characters , such as ​​ torch ​​ wait .

about UI aspect , After this update, it can be completed automatically when the query statement is entered on the page , Syntax highlighting and detection . And each time you press the Enter key, a new query will be triggered . For on page ​​/rules​​ and ​​/targets​​ The long rules and names of are also optimized , Can better query , Click on ​​/targets​​ Under the ​​Collapse All​​ Can fold all Target Click on ​​Expand All​​ You can expand all Target.

For time series database TSDB , Every time you compress, you will quickly delete the removable blocks , This can save disk space . Detailed release notes can be viewed ​ ​release notes​​ .

Reference link

• https://github.com/prometheus/prometheus/releases/tag/v2.26.0 ,By prombot of GitHub