Q: Your CPU alert threshold is 90%. The server CPU oscillates between 89% and 92% every few seconds. This causes PagerDuty to trigger the alert, resolve it, and trigger it again 50 times an hour. How do you fix this?
This is called a Flapping Alert. To fix it, you introduce Hysteresis or a Pending Duration.
for durations in PromQL.. I structure my answer around systematic triage first, root cause analysis second, and permanent remediation third.""
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: "Our SRE team tackled this monitoring and metrics bottleneck to eliminate false-positive alert fatigue. The interviewer is testing: Flapping alerts, hysteresis, for durations in PromQL.. I structure my answer around systematic triage first, root cause analysis second, and permanent remediation third."
Production Solution & Architecture
This is called a Flapping Alert. To fix it, you introduce Hysteresis or a Pending Duration. In Prometheus, this is solved using the for clause in the alert evaluation rule. Rather than firing the millisecond the CPU hits 91%, the metric must *sustainably remain* above 90% for a continuous, unbroken 5-minute window. If it drops to 89% at minute 4, the timer resets. This guarantees you are only paged for sustained actual load, completely eliminating noise from instantaneous spikes.
alert: HighCPU
expr: cpu_usage > 90
for: 5m
- Immediate Triage: This is called a Flapping Alert. To fix it, you introduce Hysteresis or a Pending Duration.
- 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 alert threshold 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 alert threshold 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 alert threshold 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 alert threshold Instant alerts triggered by transient sub-second CPU or latency blips page on-call engineers unnecessarily and dilute response urgency.