Free analyzer · No account required
Latency percentile analyzer
See the slow tail behind an average with percentiles, threshold counts and a latency distribution.
Your inputs stay in this browser tab. Saving to this device is optional.
Draft name and device saving
Current draft matches the last save or initial state.
Complete your inputs
Calculated result
Awaiting complete inputsEnter your values to see the result, or choose Try an example to explore a completed worksheet.
How this result is calculated
Sort individual request durations from fastest to slowest. The nearest-rank P95 is the value at position ceil(0.95 × sample count), counting from one. A percentile describes this sample; it does not predict the next request. The histogram groups durations into equal-width intervals to show how much of the sample sits near the slow tail.
Keep your work
Resume later
Save an editable JSON draft. Use Open draft to bring it back into this tool.
Share a report
Reports include your inputs, assumptions and any unresolved checks.
Read the complete report
# Latency percentile analyzer Tool defaults; replace with your settings before operational use. ## Review issues - Paste at least one latency value. ## Inputs Dataset and measurement window: Not entered Latency unit: ms Latency samples: Not entered Slow-request threshold: Not entered ## Method Sort individual request durations from fastest to slowest. The nearest-rank P95 is the value at position ceil(0.95 × sample count), counting from one. A percentile describes this sample; it does not predict the next request. The histogram groups durations into equal-width intervals to show how much of the sample sits near the slow tail. Use raw request durations from the same scope. Do not paste histogram bucket counts or average precomputed percentiles across hosts. Small samples produce coarse percentiles; timeouts and dropped measurements must be accounted for separately. This tool never assumes omitted requests were fast. https://aiopssre.com/latency-percentile-calculator/
How it works
Sort individual request durations from fastest to slowest. The nearest-rank P95 is the value at position ceil(0.95 × sample count), counting from one. A percentile describes this sample; it does not predict the next request. The histogram groups durations into equal-width intervals to show how much of the sample sits near the slow tail.
Follow the example
In the 20-request example, the mean is 99.9 ms, P50 is 27 ms, P95 is 300 ms and P99 is 1,000 ms. Two requests exceed 200 ms. The 1,000 ms request pulls the mean upward, while P50 still describes the middle of the sorted sample.
What to account for
Use raw request durations from the same scope. Do not paste histogram bucket counts or average precomputed percentiles across hosts. Small samples produce coarse percentiles; timeouts and dropped measurements must be accounted for separately. This tool never assumes omitted requests were fast.