Bryan Boreham
Bryan Boreham
### Proposal This could reduce memory usage. Currently, both the [series cache](https://github.com/prometheus/prometheus/blob/5f9f01de1451561a4f63a9a6171d3f591e4f9e61/scrape/scrape.go#L969C2-L969C8) and the [dropped series cache](https://github.com/prometheus/prometheus/blob/5f9f01de1451561a4f63a9a6171d3f591e4f9e61/scrape/scrape.go#L973) are implemented as Go maps with the scraped metrics string as the key....
This code: https://github.com/prometheus/client_golang/blob/9358c16714047a9913e13285f8d0f8b21e2c80b1/prometheus/registry.go#L452-L456 will end the whole program if the collector panics, since it is run on a background goroutine. This may be an unpleasant lesson. As a comparison, the...
#### What this PR does This one call to io.ReadAll was generating 6% of all garbage in some production store-gateways recently. ~Since gzip puts the size at the end of...
Currently, CPU metrics for Kubelet, Containerd, etc. are dropped by a relabel rule since they don't have a container image. On most systems they run in their own cgroup called...