Ceph » mgr » Dashboard

Problem:
When an OSD reaches nearfull or full status, the cluster health changes to WARN or ERROR, but from the landing page there's no other sign of failure:

• The OSDs card shows 100% up and running without any warning or hint.
• If you go to the OSDs page, there's no warning either (only by looking at the usage/progress bars you can guess which ones are).
• No alerts or asyncronous notifications pop up.

Improvements:
1. The landing page should display in the OSDs card at least how many OSDs are on full or nearfull statuses.

2. The landing page Raw Capacity chart should turn yellow/orange after some threshold (cluster-wide near-full ratio) and red (full-ratio).

3. Every usage bar (progress) of the OSD page should change color when hitting nearfull ratio (from blue to yellow or orange) and to red when in full ratio.