I've deployed the exporter in a Ubuntu 22.04 machine running Docker version 20.10.22, build 3a2c30b and the latest exporter version. We are using wildcards to grab everything from the project:
Around 500 projects
Around 13k refs (just main, master and tags)
Our metrics are growing exponentially - that is probably due to bad configuration for sure - but the exporter is behaving in an unexpected manner:
As we can see, it stops comunicating with Prometheus, but still consumes CPU and RAM. I would expect that under a crashing condition - also logs don't reveal anything - the process would stop, restarting the container.
Also, given your experience, for this amount of refs, does it warrant the HA deployment instead ?
Hi,
I've deployed the exporter in a Ubuntu 22.04 machine running Docker version 20.10.22, build 3a2c30b and the latest exporter version. We are using wildcards to grab everything from the project:
Our metrics are growing exponentially - that is probably due to bad configuration for sure - but the exporter is behaving in an unexpected manner:
As we can see, it stops comunicating with Prometheus, but still consumes CPU and RAM. I would expect that under a crashing condition - also logs don't reveal anything - the process would stop, restarting the container.
Also, given your experience, for this amount of refs, does it warrant the HA deployment instead ?