https://github.com/prashant-shahi/prometheus-benchmark

https://github.com/prashant-shahi/prometheus-benchmark

Last synced: 26 days ago
JSON representation

• Host: GitHub
• URL: https://github.com/prashant-shahi/prometheus-benchmark
• Owner: prashant-shahi
• License: mit
• Created: 2022-03-10T11:27:04.000Z (almost 3 years ago)
• Default Branch: main
• Last Pushed: 2022-03-10T11:27:28.000Z (almost 3 years ago)
• Last Synced: 2024-10-04T16:42:35.166Z (3 months ago)
• Language: Smarty
• Size: 24.4 KB
• Stars: 0
• Watchers: 3
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Prometheus benchmark

Prometheus benchmark helm chart is used for deploying a simple

benchmark setup to k8s cluster for benchmarking read (executing PromQL/MetricsQL queries)

and write (via [Remote Write](https://prometheus.io/docs/prometheus/latest/configuration/configuration/#remote_write) protocol)

path of Prometheus-compatible TSDBs. 

In VictoriaMetrics we use it to run tests against 

our cloud solution, compare performance and compression between releases.

The helm chart deploys three pods:

* nodeexporter + nginx (2 containers in one pod), where nodeexporter used as a metrics source

and nginx as cache-server to reduce pressure on nodeexporter;

* vmagent to scrape nodeexporter metrics and forward via remote-write to configured destinations;

* vmalert + alertmanager (2 containers in one pod), where vmalert executes 

[alerting rules](files/alerts.yaml) and sends notifications to alertmanager. Alertmanager is configured 

to blackhole received notifications. vmalert + alertmanager pod is optional and used for generating the 

read load. To disable pod creation set `.Values.vmalert.enabled=false`.

Please, check [values.yaml](values.yaml) for configuration params.

## Articles

[Benchmarking Prometheus-compatible time series databases](https://victoriametrics.com/blog/remote-write-benchmark/).

## How to run

Check [Makefile](Makefile) for make commands.

Please check the `NAMESPACE` and `RELEASE_NAME` variables in [Makefile](Makefile)

before applying any commands.

```bash

make install # to install the chart

make delete # to delete the chart

```

Please note `resources` section [values.yaml](values.yaml) and adjust it accordingly to 

configured workload. The most of resources are supposed to be consumed by vmagent

and nginx+nodeexporter pods.

## Monitoring

vmagent is configured to scrape and send its own metrics 

with job label `vmagent`. These metrics will be written to the

configured `.Values.vmagent.url.remoteWrite` destination.

Use [grafana dashboard](https://grafana.com/grafana/dashboards/12683)

to monitor vmagent's state.

vmagent is not aware of `url.remoteWrite` VictoriaMetrics configuration 

or its components, so it can't scrape their metrics. Please, configure 

monitoring of `remoteWrite` destination manually by setting up an external monitoring 

or updating [configmap.yaml](templates/vmagent/configmap.yaml) with corresponding

targets. Use grafana dashboards for [single](https://grafana.com/grafana/dashboards/10229)

or [cluster](https://grafana.com/grafana/dashboards/11176) versions of VictoriaMetrics.