Time Series and Temporal Context
You'll learn how to choose and configure time ranges to make your observability dashboards truly insightful.
Why Time Context Matters
When monitoring infrastructure, the when is just as important as the what. A metric showing "95% CPU usage" means entirely different things if it happened once for 10 seconds versus continuously for an hour. Time series data—measurements collected over time—forms the backbone of observability, but its value depends entirely on how you present temporal context.
Designing Time-Based Visualizations
Think of time like a camera zoom lens: get too close and you see noise; zoom too far out and you miss critical details.
Appropriate Time Ranges
Match your time window to your question:
- Short windows (5-15 minutes): Real-time troubleshooting, catching spikes
- Medium windows (1-24 hours): Daily patterns, incident investigation
- Long windows (days to months): Trend analysis, capacity planning
Temporal Comparison Features
Your dashboard should help users answer "is this normal?" Include:
- Week-over-week overlays: Compare current Monday to last Monday
- Day/night patterns: Does traffic always spike at 9 AM?
- Before/after views: Show the state pre- and post-deployment
Time Alignment
When displaying multiple metrics, ensure they share the same time axis. Misaligned timelines make correlation impossible—like trying to sync two videos that started at different moments.
Refresh Rates
Balance freshness with system load. Live dashboards showing SLOs might refresh every 30 seconds, while capacity planning views can update hourly.
Key Takeaway: Effective time-based visualizations require choosing appropriate time ranges for the question being answered, providing temporal comparison features to establish context, and ensuring consistent time alignment across all displayed metrics.