What are the metric monitoring tools used for ipfs-cluster?

[vasa-develop]

I went through the ipfs-cluster code and issues, but I couldn’t find the exposed endpoints for gathering the metrics.
Seems like we are using Prometheus/Opencensus for monitoring (please do tell me if I am wrong), but I am not sure about it.

Questions:

• What monitoring tools are supported right now?
• What are the exposed endpoints for the same?
• Do we have Grafana board JSON for IPFS Cluster? I am actually planning to create advanced boards for managing the cluster. So, if someone is already working on it, I would be happy to join in

[hector]

hey, there are some docs here: https://cluster.ipfs.io/documentation/configuration/#the-observations-section . We support prometheus for metrics and jaeger for tracing and the endpoints are configured there.

We don’t have a Grafana board, but I think it would be great to have one.

@lanzafame knows all the things about monitoring so he can help!

[vasa-develop]

Sweet. I got the metrics!
@lanzafame As I am creating Grafana board for IPFS Cluster, do you have any inputs on what metrics should I include in it?

[postables]

I’ve got a bunch of tools that i already use to monitor IPFS Cluster with grafan and zabbix, feel free to poke around:

• https://github.com/RTradeLtd/Temporal/tree/master/setup/configs/grafana (grafana dashboards)
• https://github.com/RTradeLtd/Temporal/blob/master/setup/scripts/ipfs/ipfs_cluster_monitor.sh (script to use with zabbix)
• https://github.com/RTradeLtd/Temporal/blob/master/setup/configs/zabbix/userparameters/userparameter_ipfs.conf (zabbix configurations)

[HamidullahMuslih]

I added below metrics endpoint as a prometheus scrap job.

http://localhost:5001/debug/metrics/prometheus

but I could not find IPFS dashboard in Grafana https://grafana.com/grafana/dashboards/.

Is there anyone who can help me, or I should build the dashboard myself?

[forgondolin]

Yep, the grafana dashboards should be created by you.
There are some examples on internet:

https://gist.github.com/eefahy/52c3df180e67a2fc4520a7be1dc394b7

[hector]

ipfs-cluster-service exposes its own prometheus endpoint, it needs to be enabled in the configuration or doing ipfs-cluster-service daemon --stats and it’s by default on port 8888 (Configuration - Pinset orchestration for IPFS).