# Conflicts:
#   alerts/db-under-heavy-load.rules
#   alerts/git-cat-file.rules
#   alerts/postgresql-replication-status.rules

Adjust `FOR` interval on alerts to avoid scrape failure induced alert flapping.

* Monitor ext* and xfs, not NFS filesystem mounts.
* Print filesystem alerts with a percentage, not the raw float.
* Extend SSL cert alerting condition to 30min to avoid lost probes
  triggering alert flapping.

- Changed all alert names to CamelCase (@bjk-gitlab asked for this)
- Improved/fixed some alert conditions and messages
- Try to preserve a tad more labels in expressions (when possible)

Unfortunately the Slack notification templates in Alertmanager rely on
there being only one "title" and "description" value, because they use
the title/description from CommonAnnotations (so those are only present
if they are the same between all alerts resulting from one alerting
rule). So I kept appropriate aggregations and topk-ing in some of the
alerts to not break that assumption.

Obvious typo is obvious, isn't it?

Signed-off-by: Yongmin Hong <revi@pobox.com>