Upgrading debian causes issue on tsdb compaction chunk size

Hello everyone,

I upgraded my prometheus servers and since the upgrade, I encountered a problem on a rule that calculates the remaining disk space (based on my retention). It uses prometheus_tsdb_compaction_chunk_size_bytes_sum

I have an increase of this variable without having any change in the infrastructure (no more scrapping / metrics / …).

Here is the graph that shows the unexpected increase:

Is it “normal”?

Thank you :slight_smile: