Q: Your Prometheus Time-Series Database (TSDB) is running on a massive disk with plenty of space left, but it is thrashing the CPU and IOPS with high "Compaction" activity. What causes excessive compaction?
High TSDB compaction (and resulting IOPS thrashing) is heavily correlated with Metric Churn.
Want to master this scenario in a live sandbox? The Linux Foundation's Prometheus Certified Associate (PCA) & Monitoring Labs covers this exact problem with hands-on terminal drills.
🛠️ Production Runbook & Step-by-Step Resolution
Direct Answer: "Logs tell you what happened, metrics tell you where to look, and distributed traces pinpoint the exact slow component. The interviewer is testing: Metric churn, TSDB internal mechanics, head block vs persistent blocks.. I structure my answer around systematic triage first, root cause analysis second, and permanent remediation third."
Production Solution & Architecture
High TSDB compaction (and resulting IOPS thrashing) is heavily correlated with Metric Churn. Churn is different from High Cardinality. Churn happens when a service creates brand new metric series, stops updating them after highly ephemeral periods, and creates new ones. For example, if a developer mistakenly uses the Kubernetes Pod IP as a metric label in a rapidly auto-scaling environment. Every time a pod is replaced, the old metric series is abandoned, and a new one is created. Prometheus groups recent data in temporary memory blocks. When moving to persistent disk, it "compacts" related series. Massive churn forces the compactor to constantly rewrite indices and stitch together millions of fragmented, short-lived series, burning massive CPU. The fix is to remove ephemeral labels (like Pod IPs) and use static identifiers (like Service Names).
- Immediate Triage: High TSDB compaction (and resulting IOPS thrashing) is heavily correlated with Metric Churn.
- Run targeted verification commands before modifying configuration.
- Automate permanent guardrails (CI check, alerts, IaC policy) to prevent recurrence.
Senior hiring managers and bar raisers evaluate how you prioritize during outages. Avoid these common anti-patterns:
- Injecting high-cardinality label dimensions (such as user IDs or request UUIDs) into Prometheus metrics when configuring Observability Prometheus Time High-cardinality label values cause exponential time-series growth, consuming gigabytes of TSDB memory and crashing Prometheus servers.
- Evaluating application performance solely through arithmetic mean or p50 latency while ignoring tail latency (p99) when configuring Observability Prometheus Time Averages hide catastrophic tail latency outages where 1% of high-value user transactions suffer severe timeouts.
- Failing to inject and extract W3C distributed trace context across message brokers and RPC calls when configuring Observability Prometheus Time Breaking distributed trace context propagation fragments end-to-end transaction waterfalls into disconnected, un-correlatable trace segments.
- Configuring alert rule evaluation intervals without sufficient duration damping (for: 5m), inducing alert flap when configuring Observability Prometheus Time Instant alerts triggered by transient sub-second CPU or latency blips page on-call engineers unnecessarily and dilute response urgency.