vespa-engine / vespa_exporter Goto Github PK

Vespa exposes a lot of metrics, and this exporter collects them from all the nodes of a cluster. So when it exposes them it can generate a big quantity of data (I'm experiencing a size of 65Mo). It is not very suitable since the Prometheus server will thus take a lot of time to load the data exposed by this exporter. And I only use a few of those metrics.

A simple solution would be to give the ability to add a configuration file to this exporter with the names (in Prometheus format) of the metrics it has to expose. So that the users of the exporter could choose to expose only the metrics they need.

What do you think about that ?

I would gladly implement a solution to this issue.