Prometheus operators can't keep important metrics longer than unimportant ones.
Teams running Prometheus who pay to store metrics they rarely need
Where it stands
- Sourced (done)
- Kit (done)
- Wanted (next)
- Building
- Solved
How this problem came to be
- Found2 sourcesgithub.com/prometheus
- Quoted4 claimseach checked against its page
- BriefWrittenand what solved looks like
- Market1 signals20 reactions on the issue
- Kit5 steps21 tools · 6 routes
- People0 want0 live it · 0 would test
The problem, broken down
- Who it hurts
- People who run Prometheus monitoring and manage its disk use.
- The moment
- When storage fills up and the only choice is to shorten retention for everything.
- What people do today
- One retention time applies to all metrics.
- Why it's unsolved
- It touches how the time-series database compacts its blocks, and the scope is still being discussed.
- What solved looks like
- A prototype in Prometheus's storage compaction that drops series matching a configured selector once they pass their own retention, with a benchmark of disk saved on a sample dataset and written answers to the scope questions raised in the issue.
Evidence
Every claim is tied to the source it comes from- 1The issue says Prometheus applies the same retention time to all metrics, so less important metrics take up storage as long as critical ones.
- 2It asks for retention per metric, namespace or label, set in the config file or through the API.
- 3A commenter pointed to earlier discussion and suggested that dropping some series during compaction would be simpler.
- 4In 2026 a contributor traced the compaction path and asked whether expired series should be dropped whole or trimmed.
Builders still need to know
Written from the cited sources. Anything they don't say is left as a question, not a guess.
Sources
From a public page with a citation.
What the source says, in short
The issue says Prometheus applies the same retention time to all metrics, so less important metrics take up storage as long as critical ones. It asks for retention per metric, namespace or label, set in the config file or through the API. A commenter pointed to earlier discussion and suggested that dropping some series during compaction would be simpler. In 2026 a contributor traced the compaction path and asked whether expired series should be dropped whole or trimmed.
Forks 0
Nobody has forked this yet. Be the first.
Market read
How big this is, from the evidence
- GitHub ↗Builders' pull · Medium confidence20 reactions on the issue
How these were measured: Total reactions on the GitHub issue, read from its API.
For whoever takes this up
The builder's path for this problem
Five steps, each with the tools that fit this problem and a word to carry with you. A starter prompt is ready to paste into any assistant.
Open the pathDiscussion 0
No comments yet.