error rate

The share of requests that fail, reported alongside latency so a fast result is not mistaken for a healthy one.

In more detail

A service that fails quickly can have excellent latency, so always report count, errors and latency together. Count timeouts and wrong answers as errors, not only crashes.

Examples

The rollback rule fires when the canary’s error rate goes above 1%, for example 4.8%.
sum(rate(http_requests_total{status=~"5.."}[5m])) / sum(rate(http_requests_total[5m]))

Related terms

Taught in

References

Certification courses