fix many-to-many not allowed in Grafana dashboard

[p-se]

One part of the issue is that the metrics from the ceph mgr contain labels for instance, though, the instance label is also automatically populated by Prometheus. When Prometheus does that, the instance label becomes the hostname of the Ceph manager node, which is wrong. By setting honor_labels in the scrape configuration of ceph-mgr to true, the instances labels exported by the Prometheus manager module are preserved.

Partly fixes bsc#1175585

Signed-off-by: Patrick Seidensal pseidensal@suse.com

---

Checklist:

• [x] Added unittests and or functional tests
• [ ] Adapted documentation
• [x] Referenced issues or internal bugtracker
• [x] Ran integration tests successfully (trigger with "@susebot run teuthology" in a GitHub comment; see the wiki for more information)

[jschmid1]

@susebot run teuthology

[susebot]

Commit a668f2e6a5939c4fe1ba6a01f0b6fac082236a17 is OK for suite deepsea:tier2. Check tests results in the Jenkins job: https://storage-ci.suse.de/job/pr-deepsea/467/