Prometheus and ClickHouse handle high-cardinality metrics through fundamentally different architectures. Prometheus pays cardinality costs at write time through memory overhead, index maintenance, and series creation (~3-4KB per active series), leading to OOM failures during ingestion. ClickHouse defers costs to query time,

14m read timeFrom last9.io
Post cover image
Table of contents
Why This MattersThe War StoryA Note on FairnessThe Core Question: Where Does Identity Live?Prometheus: Write-Time IdentityPrometheus Internals: Why Cardinality Hurts Early"Just Add More Prometheus" Doesn't WorkNative Histograms: A Schema-Level SolutionClickHouse Internals: Why Cardinality Feels DifferentThe ComparisonWhy Hybrid Pipelines ExistWhat About TimescaleDB, QuestDB, InfluxDB, Druid?The Design-Time DecisionClosing Thought

Sort: