While bringing up Prometheus 2.0 for Scalefastr (our Cassandra and Elasticsearch managed hosting platform) we noticed that a lot of the Grafana dashboards for Cassandra, Elasticsearch, etc worked acceptably on Prometheus 1.x but did not migrate cleanly to 2.x.
Additionally, we had some problems with unified labels across dashboards as well as the fact that many dashboards were not well documented.
We ended up having to reverse engineer what the original dashboard author meant as well as trying to reproduce their Prometheus configuration (which wasn't always clear).
To help others migrate to Prometheus 2.0 (as well as have functioning dashboards) we thought it would be valuable to have a "known good" configuration of both dashboards and Prometheus configuration that they would work from.
You should be able to take the raw .json and paste it into Grafana for dashboard import. You can also use the Grafana API to import the dashboards from JSON.
You will have to set your datasource or use a default datasource.
If you make any changes you can use:
sbin/fetch-dashboards.sh
to export your dashboards as we would love to integrate any valuable changes you've made.
Feel free to send a pull request if you make any changes.
The prometheus.yml
file we include is our full Prometheus configuration
with hosts blinded. This will specify all the exporters we use as well as
the labels applied.
While developing these we didn't track all the original authors of the dashboards.
At the time we weren't planning on creating this project and just assumed that these would be internal dashboards. If we didn't give you credit please let us know and we will rectify. It's our plan to make sure everyone gets credit
The Cassandra metrics are based on Cassability: