Open dsvensson opened 6 years ago
Curious about why not more common metric names are being used, like
http_request_duration_milliseconds{job=..., path=/foo/bar, status=200, method="get"} (histogram) http_request_size_bytes{job=..., path=/foo/bar, status=200, method="post"} (histogram) http_request_total{job=..., path=/foo/bar, status=403, method="delete"} (counter) http_requests_total{job=...} (counter)
It seems to be more aligned with: https://prometheus.io/docs/practices/naming/
...and in general how other HTTP services use Prometheus. So a much bigger focus on labels rather than creating unique metrics.
Curious about why not more common metric names are being used, like
It seems to be more aligned with: https://prometheus.io/docs/practices/naming/
...and in general how other HTTP services use Prometheus. So a much bigger focus on labels rather than creating unique metrics.