lhotari opened a new pull request, #592:
URL: https://github.com/apache/pulsar-helm-chart/pull/592

   ### Motivation
   
   kube-prometheus-stack has been replaced by victoria-metrics-k8s-stack for 
Pulsar Helm Chart version 4.0.0 due to
   Prometheus incompatibility with Pulsar metrics since Pulsar 2.11.0 . Pulsar 
exposes metrics in a format that is partially OpenMetrics 1.0.0 text format, 
but isn't fully compatible. Prometheus doesn't provide proper support for 
OpenMetrics 1.0.0 text format, even in Prometheus version 3.2.1 where it was 
extensively tested before making the switch to Victoria MetricsVictoria Metrics 
is Apache 2.0 Licensed OSS and it's fully compatible with Prometheus.
   
   ### Modifications
   
   - replace kube-prometheus-stack with [victoria-metrics-k8s-stack 
](https://github.com/VictoriaMetrics/helm-charts/blob/master/charts/victoria-metrics-k8s-stack/)
 which is Apache 2.0 Licensed OSS and it's fully compatible with Prometheus.
   - The Pulsar Helm Chart has pre-installed Grafana dashboards for Pulsar from 
various Apache 2.0 Licensed sources. These dashboards are outdated, and there's 
a separate project https://github.com/lhotari/pulsar-grafana-dashboards where 
the dashboards will be cleaned up and made fully compatible with Pulsar 4.0+ 
and Victoria Metrics.
     - In Pulsar, the Prometheus client_java library was updated to 0.15.0 with 
https://github.com/apache/pulsar/pull/13785 in Pulsar 2.11.0 and this broke 
most of the metrics since the output is not Prometheus Text 0.0.4 format any 
more. 
     - Since Victoria Metrics has better support for handling scraped metrics, 
it should be possible to get the metric names working again with Victoria 
Metrics configuration. In Prometheus, this isn't possible at the moment. There 
are some changes going on in the Prometheus project, to address it, but since 
Victoria Metrics is more feature complete and runs with less resources, it's a 
better choice in getting the dashboards usable again.
   
   ### Additional context
   
   - Another PR will contain changes to migrate from using Prometheus Operator 
CRDs to use Victoria Metrics CRDs (PodMonitor -> VMPodScrape)


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscr...@pulsar.apache.org

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org

Reply via email to