Frederic Branczyk
Frederic Branczyk
The heuristic of allowing inserts up to 0.5x timespan of head blocks is based on the assumption that we can safely and correctly cut blocks at that time, I'm wondering...
There have been discussions on per time-series retention on upstream Prometheus before, I think at least having a discussion with the team is worth it, just to see if there...
Configuration is a technicality, I'm not entirely sure relabelling would work exactly but something close to that probably yes. I agree the compactor is the component that needs to take...
This is being worked on on prometheus, once done there I would say we can implement it here with the same semantics and configuration.
If this is something you need to do, then I'd prefer if you deploy a proxy that has hostNetwork that forwards the request to etcd instead of exposing a potentially...
I think there is space for something like this in some shape or form in the prometheus remote write protocol. Essentially the protocol itself should have the ability to even...
There are some caveats: 1) The receive component will only ingest data up to 1 hour old. 2) Prometheus will stop retrying sending data once it's not available in the...
1. The heuristic is 0.5x of the smallest block produced and since the default smallest block is 2h, the heuristic determines 1h. Further explained: this is because the WAL accumulates...
Just in case this is not already known, we recently introduced on parca server side to throw away any address that can never be symbolized, eg has no mapping. We...