Cloudwatch Exporter configuration for existing Grafana Dashboards

[Uğurcan Çaykara]

Hello everyone,

I have currently have an eks cluster running on AWS. I installed helm chart to setup prometheus and grafana. All metrics I need for EKS(deployments, services, pods etc.) are totally fine. So I wanted to centrally use that grafana for other AWS services that I use and that's why I configured config.yml for lambda metrics as given at the repository and deployed cloudwatch exporter (https://github.com/prometheus/cloudwatch_exporter/). I can see the related metrics at the Grafana dashboard. When I hit the explore tab at left menu from Grafana UI and enter lambda, all metrics given at the config.yml for lambda metrics are totally fine. I can query them. And now I wanted to use a dashboard for lambda -> https://grafana.com/grafana/dashboards/593-aws-lambda/ 

However it uses cloudwatch as datasource not prometheus and that's why I can't see no data for that specific dashboard. What's best way to overcome this. Is it something quickly editable and fixable ? or is it better to start creating dashboard from beginning. If someone. can help me here, I would appreciate that. 

Thanks

Best regards

[Stuart Clark]

If you use dashboards from the Grafana site that aren't designed for
Prometheus their usefulness is limited. Other datasources (such as
Cloudwatch or Datadog) use totally different query languages, so what
you are really gaining is an outline of a design rather than anything
you can directly use. You would need to rewrite all the queries to
operate in a similar manner using PromQL.

Whenever I'm having a look for Grafana dashboards I use the datasource
filter such that only those designed specifically for Prometheus are
listed - at least then they mostly work out of the box (although often
still need slight tweaks due to different job names or use of labels).

--
Stuart Clark