Watching server disk space from the Server Versions page

The disk-space column on the Server Versions page shows the percentage of disk used on each server in your cluster — and it is the fastest way to catch a box that is about to run out of room before recordings stop and the dialer stalls.

What eats disk on a VICIdial server

Three categories of data dominate disk consumption on a production VICIdial box:

• Call recordings — each recorded call produces one or more audio files. A busy dialer running Stereo recording writes two-channel files that are roughly twice the size of mono. A 50-seat predictive campaign can produce several gigabytes per day.
• Asterisk and VICIdial log files — the dialer writes verbose logs by default. These rotate, but if rotation is misconfigured the logs directory can grow unchecked over weeks.
• Archives and snapshots — a Server snapshot taken before a major upgrade, left on the local filesystem, can occupy tens of gigabytes that never get cleaned up.

Any of these can grow gradually and invisibly. The disk-space column makes the trend visible without requiring a login to the server itself.

What happens when disk fills up

A VICIdial server that runs out of disk space fails in ways that are not always obvious at first:

• Call recording stops silently. The campaign keeps dialing and agents keep talking, but no audio files are written. You discover the gap only when someone tries to pull a recording and finds nothing there.
• The Asterisk process and the VICIdial perl scripts write to disk constantly. When the filesystem is full, writes fail and processes can wedge in unexpected states — calls stuck in ringing, agents unable to disposition, or the real-time display freezing.
• Database writes can also fail if the MySQL data directory shares the same partition as the rest of the filesystem. This is a hard dialer stop — no new calls, no status updates.

stateDiagram-v2
  [*] --> Normal : disk below 70 pct
  Normal --> Warning : disk crosses 80 pct
  Warning --> Critical : disk crosses 90 pct
  Critical --> RecordingStop : filesystem full
  RecordingStop --> DialerWedge : write errors on log and DB
  Warning --> Normal : old recordings purged
  Critical --> Normal : old recordings purged

Reading the column as an early-warning signal

A reasonable threshold to work with: below 70% is comfortable, 70–80% warrants a note on the calendar to schedule a cleanup, above 80% should trigger action before the next heavy dialing day, and 90% is a near-emergency. The exact numbers depend on how fast your campaign fills disk — a single Campaign running stereo on a 200-seat floor fills disk far faster than a smaller team.

Common remediation steps when the column reads high: move old recording files to remote storage, check whether your Recording retention policy is actually purging files on schedule, and remove any leftover snapshot archives from the local filesystem. After cleanup, reload the Server Versions page to confirm the percentage has dropped.

Fitting disk monitoring into your routine

Checking the disk column takes about three seconds when you load the Server Versions page. Most operators make it part of the same morning check that confirms load and active status. For a time-series view of how the server is trending overall, see the Server Performance Report — it covers the load and channel trends that pair with this disk check.

Disk management is one component of a broader server-health habit. For the full picture, read the guide to monitoring VICIdial server health and capacity.

If managing disk retention on a self-hosted box is not how you want to spend your time, start a VICIfast trial — your managed server is up in under 40 seconds, with automated retention policies configured from the start.