Bugfix - handle connection failure behavior

sapcc / nsx-t-exporter

Prometheus NSX-T exporter

Apache License 2.0

14 stars 10 forks source link

Bugfix - handle connection failure behavior #5

Closed tsimchev closed 4 years ago

tsimchev commented 4 years ago

do not update cache on connection failure
add additional metric about last successful scrap of NSX-T API
- this metric will be updated only when the collection from all of the NSX-T API endpoints is successful.

joker-at-work commented 4 years ago

do not update cache on connection failure

Does that mean we report stale metrics? And the only way to figure this out is by finding the additional metric about the last successful scrape?

tsimchev commented 4 years ago

do not update cache on connection failure

Does that mean we report stale metrics? And the only way to figure this out is by finding the additional metric about the last successful scrape?

yes, this is the approach we discussed yesterday.

Some Additional Options that we discussed

add additional value -999 to metrics, indicating conn failure or time out
hide metric from the list
add an additional metric for every metric ending with _since in seconds

joker-at-work commented 4 years ago

OK, if that was the consent.