Monitoring

Docker Manager watches the disks of every server: their own health reports (SMART) and their RAID arrays, so you can replace a disk before it fails. A disk or array with a problem raises an alert. For CPU, memory, network and disk use over time, see Metrics.

Check a server's disksCopy Anchor Link

1. Open Environments, pick an environment and open its System tab.
2. The Disk Health card lists each disk with its Health, Temperature, how long it has been Powered On and its Issues in words, for example "8 reallocated sectors" or "Worn 92%". Serial numbers and firmware are under Advanced.
3. Press the info button (ⓘ) at the end of a disk's row to see all its SMART data: every attribute with its value, worst, threshold and raw value, and the other values the disk reports, such as power cycles or data written. Values that matter for the disk's health are marked OK, warning or problem in their Status column.
4. Press Check Disks Now to read every disk again right away. The agent also reads them every 30 minutes; Last checked says when it last did.

A disk's health is one of:

• Healthy: nothing to do.
• Warning: the disk reports damaged or pending sectors, errors, wear of 90% or more, or little spare capacity. Keep your backups current and plan a replacement. A disk that is, or once was, hotter than its own limit also shows Warning: check the server's cooling. The disk keeps that record, so the warning stays after it cools down.
• Failing: the disk's own self-check failed or it reports a critical problem. Back up and replace it soon.
• Sleeping: the disk was in standby and wasn't woken. It shows the values of its last check, without temperature. A disk whose last check found a problem keeps showing Warning or Failing while it sleeps. A disk that sleeps through every check is woken for one once a day.
• Unreadable: the disk could not be read, didn't answer in time or reported no health data. Its Issues say why.
• Missing: the agent found the disk before but no longer does. It may have failed or been removed. It stays in the list until it is found again or the agent restarts.

Check RAID arraysCopy Anchor Link

If the server has Linux software RAID or ZFS pools:

1. On the server's System tab, the RAID card lists each array (a software RAID array by its path, such as /dev/md0) with its state (Healthy, Degraded, Rebuilding, Checking, Failed or Inactive), its disks (a failed one in red) and the progress of a rebuild or check with the time left.
2. For a software RAID array, press the info button (ⓘ) to see its level, size, metadata, chunk size, layout and bitmap, and each member with its slot, role and the health of its disk.
3. Press Check RAID Now to read the arrays again.

A server without arrays shows No RAID arrays found in the Disk Health card.

Give the agent access to the disksCopy Anchor Link

To read the disks, the agent runs with privileged: true. The Quickstart files and the install commands do this. If the Disk Health card says No Disk Access, add the line to that server's agent:

1. Open Stacks and the server's docker-manager stack, then its Files tab.

2. In compose.yaml, add privileged: true to the docker-agent service:

   Copy Text

   services:
     docker-agent:
       # ...keep the existing lines...
       privileged: true

3. Save the file and press Deploy.

If you haven't imported Docker Manager's project yet, edit the compose.yaml on the server and run docker compose up -d there.

Good to know:

• Checks only read the disks: they never start a disk self-test or a RAID scrub.
• Disks you connect later appear after the agent restarts.
• Virtual disks have no health data; the card says so. btrfs RAID and hardware RAID controllers are not covered.
• To change how often disks are checked, or to turn checks off, see Configuration.
• To chart another disk's capacity, add a read-only mount under /disks in Metrics.