Add Prometheus and Grafana to the pipeline#17
Conversation
Which is the standard port for Prometheus metrics.
6f2d2b5 to
4446616
Compare
|
Everything seems to be in place, but I can't use the available dashboards in https://github.com/0xB10C/peer-observer/tree/master/tools/metrics/dashboards. Still working to be them right. Reminder to self: include documentation about the http endpoints we get with this pipeline. |
|
Added up-to-date "playlist" dashboards in peer-observer/peer-observer#213 along with some documentation in https://github.com/0xB10C/peer-observer/tree/master/tools/metrics#grafana-dashboards - let me know if you run into any problems. |
| scrape_configs: | ||
| - job_name: 'peer-observer-metrics' | ||
| static_configs: | ||
| - targets: ['metrics:8001'] |
There was a problem hiding this comment.
To allow for a multi-node setup, I use the label host in my set up for the dashboards. Even if you just have one host, you probably want to add a name for it here. See also https://github.com/0xB10C/peer-observer/tree/master/tools/metrics#grafana-dashboards
| - targets: ['metrics:8001'] | |
| - targets: ['metrics:8001'] | |
| labels: | |
| host: node1 |
grafana/dashboards/messages.json
Outdated
| @@ -0,0 +1,365 @@ | |||
| { | |||
There was a problem hiding this comment.
if possible, it might be better to provision the dashboards from the peer-observer repo directly and not maintain them here too
|
This might be helpful: You can do this similar to https://github.com/0xB10C/peer-observer-infra-library/blob/e14fc1e29a3adbd35d5e3af73766039279fbf5c4/modules/web/web.nix#L451-L466 (which is in Nix, but you get the idea). I don't think the infrastructure focused dashboards in https://github.com/0xB10C/peer-observer-infra-library/tree/master/modules/web/dashboards are too relevant (since you don't run the node-exporter in the container I think), but you might be able to come up with others that are relevant for this docker setup. |
No description provided.