Q: After enabling service mesh telemetry, Prometheus cardinality explodes because metrics include source pod, destination pod, path, method, response code, and workload labels. How do you control it?
Service mesh telemetry is powerful but can create a series for every source-destination-path combination.
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: Service mesh metrics, label control, aggregation strategy.. I structure my answer around systematic triage first, root cause analysis second, and permanent remediation third."
Initial Diagnostics & Root Cause Analysis
Service mesh telemetry is powerful but can create a series for every source-destination-path combination.
- Dropping pod-level labels from high-volume metrics and keeping workload, namespace, and service labels.
- Normalizing paths, such as
/orders/{id}instead of/orders/12345. - Keeping method and response-code class, but avoiding unnecessary headers or user-level labels.
- Creating recording rules for common service-to-service RED metrics.
Remediation & Permanent Safeguards
I would control it by: The goal is service-level observability, not a unique time series for every request shape.
- Applying metric relabeling at scrape time to remove labels that are not used in alerts or dashboards.
- Setting cardinality budgets per team and reviewing top series regularly.
- Dropping pod-level labels from high-volume metrics and keeping workload, namespace, and service l...
- Normalizing paths, such as /orders/{id} instead of /orders/12345.
- Keeping method and response-code class, but avoiding unnecessary headers or user-level labels.
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 enabling service 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 enabling service 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 enabling service 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 enabling service Instant alerts triggered by transient sub-second CPU or latency blips page on-call engineers unnecessarily and dilute response urgency.