Make scrape timeout configurable
Bug #1867180 reported by
Peter Sabaini
This bug affects 1 person
| Affects | Status | Importance | Assigned to | Milestone | |
|---|---|---|---|---|---|
| Prometheus2 charm |
Fix Released
|
High
|
Peter Sabaini | ||
Bug Description
We're currently not configuring the Prom scrape_timeout, leaving it at the global default of 10s
This can be tight for highly loaded clouds (for instance with the prom openstack exporter), which will lead to silently dropped metrics.
I believe we should therefore configure scrape_timeout.
One simple heuristic would be to just set scrape_
Alternatively, provide a separate knob scrape-timeout and check for that constraint during config-changed
Related branches
~peter-sabaini/charm-prometheus2:bug/1867180
- Paul Goins: Approve
- Adam Dyess (community): Approve
-
Diff: 302 lines (+104/-56)14 files modified.gitignore (+1/-0)
Makefile (+14/-6)
config.yaml (+4/-0)
dev/null (+0/-47)
reactive/prometheus.py (+1/-0)
templates/prometheus.yml.j2 (+1/-1)
tests/bundles/bionic.yaml (+8/-0)
tests/bundles/overlays/bionic.yaml.j2 (+1/-0)
tests/bundles/overlays/xenial.yaml.j2 (+3/-0)
tests/bundles/xenial.yaml (+8/-0)
tests/test_requirements.txt (+2/-0)
tests/tests.yaml (+11/-0)
tests/tests_prometheus.py (+39/-0)
tox.ini (+11/-2)
| Changed in charm-prometheus2: | |
| importance: | Undecided → High |
| Changed in charm-prometheus2: | |
| assignee: | nobody → Peter Sabaini (peter-sabaini) |
| status: | New → Triaged |
| Changed in charm-prometheus2: | |
| status: | Triaged → In Progress |
Revision history for this message
| Andrea Ieri (aieri) wrote | #1 |
| Changed in charm-prometheus2: | |
| status: | In Progress → Fix Released |
| milestone: | none → 20.05 |
To post a comment you must log in.
available in cs:prometheus2-16