Latency Percentile Chart
API request latency as p50, p95, and p99 lines over 24 hourly points with ms axis formatting.
API request latency as p50, p95, and p99 lines over 24 hourly points with ms axis formatting.
The Application Collection unlocks the source for every Application block. All Access unlocks every Collection.
Already purchased? Log in
Latency Percentile Chart is the API status panel, three lines over 24 hourly points labeled p50, p95, and p99. The p50 holds flat between 39 and 52 ms all day while p99 spikes to 860 ms at 03:00 before recovering by 06:00. The current p99 callout reads 295 ms so the on-call reader gets the present state without scanning the right edge.
Three series share one ChartContainer with a ms tick formatter on the Y axis, and the legend distinguishes them without requiring a tooltip hover. The footnote names the cause honestly: the p99 spike between 02:00 and 04:00 lines up with the nightly export job.
Reach for this block on an API health or infrastructure dashboard, wired to your metrics backend query that emits latency percentiles per hour. Set the X axis interval to match your polling window and add a horizontal reference line if you have an SLO threshold.
A natural flow around it on an Application Pro page:
Before
After
One strong use is the API response time health panel. Other three-series line charts:
Tip: name the known spike cause in the footnote rather than leaving readers to assume the worst.