percentiles-heatmap
Input Stack:
|
⇨ | Output Stack:
|
Create a heatmap visualization from percentile distribution data by grouping on the percentile
tag.
This provides an intuitive way to visualize the distribution of values over time, with color intensity
representing the magnitude of values at different percentile levels.
Parameters¶
- expr: Time series expression for percentile metrics (must contain
percentile
tag)
Data Requirements¶
The metrics must be recorded as percentile timers or distribution summaries with the percentile
tag.
If using Spectator, use these helper classes:
- PercentileTimer - For timing measurements with percentile tracking
- PercentileDistributionSummary - For distribution measurements
How It Works¶
This operator is syntactic sugar that combines grouping and heatmap visualization:
# These expressions are equivalent:
name,requestLatency,:eq,:percentiles-heatmap
name,requestLatency,:eq,(,percentile,),:by,:heatmap
The resulting heatmap shows:
- X-axis: Time progression
- Y-axis: Percentile levels (P50, P90, P95, P99, etc.)
- Color intensity: Magnitude of values at each percentile/time combination
Examples¶
Creating a latency distribution heatmap:
Default |
name,requestLatency,:eq, :percentiles-heatmap |
Related Operations¶
- :heatmap - General heatmap visualization for any grouped data
- :percentiles - Extract specific percentile values
- :by - Group data by tag keys (used internally)
See Also¶
- Heatmap API Documentation - Additional configuration options
- Percentile Timer Pattern - How to instrument code for percentile tracking
- Percentile Distribution Summary Pattern - Distribution measurement patterns
Since: 1.8