What one row represents
One completed request before grouping by route template.
Decision this helps you make
Prioritize an endpoint for release comparison, dependency inspection, or rollback review.
Metric definitions
Put the denominator and units beside the chart
Use your own field names, but keep the same definition of one row. Check the linked schema before mapping your production events.
- Request volume
- Error rate
- p50 latency
- p95 latency
p95 ms
Demonstration data, not a customer result or benchmark
Review the query before adapting the fields
Adapt this query to your events before using it in production. Set a time range and environment filter, decide how to handle incomplete time buckets, and require enough events for a useful comparison.
SELECT
endpoint,
COUNT(*) AS requests,
ROUND(
100.0 * SUM(CASE WHEN status = 'error' THEN 1 ELSE 0 END)
/ NULLIF(COUNT(*), 0),
2
) AS error_rate_pct,
approx_percentile_cont(latency_ms, 0.50) AS p50_ms,
approx_percentile_cont(latency_ms, 0.95) AS p95_ms
FROM api_requests
GROUP BY endpoint
ORDER BY error_rate_pct DESC, requests DESC;Check these before publishing the result
- Endpoint values use route templates, such as /users/:id, instead of raw URLs.
- Latency uses one end-to-end millisecond definition across producers.
- Operational charts omit the newest incomplete time bucket.
Check the query and metric definitions
- Set a start and end time before using this query in production.
- Keep request volume beside every percentile and rate.
- Segment by release and error type after detecting a regression.
What the results mean
What this result cannot prove by itself
Always show request count beside a percentile. A high p95 from a tiny group is not equivalent to broad customer impact.
Related dashboard examples
SaaS health overview
Did account adoption, application reliability, or represented revenue change enough to require a deeper review?
Inspect exampleIncident customer impact
Which accounts and operations were directly represented in a declared incident window?
Inspect exampleLLM usage and unit economics
Which model-backed workflow consumes the most estimated cost per successful outcome?
Inspect exampleValidate the definition with known data
Test the query with known successes, failures, missing values, duplicates, and values at its limits. Check the results before using it in a production dashboard or alert.