Automatic disk resizes

Overutilizing storage on a Postgres server can be operationally dangerous because there might not be enough disk space for the server to recover in case of an emergency.

To prevent such situations, Crunchy Bridge will take actions to resize your server's disk size on your behalf. Here is a timeline of the actions that are taken as a server's disk usage grows beyond 75%:

  • 75% disk used: warn team administrators by email
  • 80% disk used: warn team administrators by email
  • 85% disk used of disks smaller than 500 GB: trigger an automatic disk resize; notify team administrators by email
  • 85% disk used of disks larger than 500 GB: warn team administrators by email
  • 90% disk used of disks larger than 500 GB: trigger an automatic disk resize; notify team administrators by email
  • Less than 1 GB left or 95% disk used out of 10GB total: put cluster into read-only mode; notify team administrators by email

Info

Note that a cluster will be put into read-only mode when disk usage becomes critical, as outlined above. This is done to protect your data while the automatic resize operation completes.

When an automatic resize is triggered, the new size is calculated based on the original size:

  • Disks smaller than 100 GB will be increased by 50% (e.g. 10 GB becomes 15 GB).
  • Disks between 100 GB and 999 GB will be increased by 25% (e.g. 100 GB becomes 125 GB).
  • Disks larger than 1000 GB will be increased by 15% (e.g. 1000 GB becomes 1150 GB).

The platform will attempt to grow the disks on your running cluster, avoiding a failover maintenance operation. This applies to the primary server as well as any HA standbys, and is significantly faster than provisioning a new server and migrating your data.

If the platform cannot complete a disk growth resize (for example, because of limits on how often disks can be modified in a short period), it falls back to the standard failover-based resize process. You can read more about this on the Cluster Management page.

Hint

Automatic resizes are an emergency action. As such, they complete as soon as they are ready, and ignore the cluster maintenance window, if any.

Manual resizing

You can trigger a manual resize by selecting "Resize" from a cluster's "Cluster Actions" menu. For details on how manual resize operates, see the Cluster resize section of the Cluster Management page.

Although Crunchy Bridge will automatically resize cluster disks for you, we strongly encourage team admins to heed the disk size warning emails and take action in advance of an automatic resize.

If you initiate a manual resize once the disk is above the 75% threshold, the system will automatically resize it by 15% or 25%, as outlined above.

Please reach out to support if you plan to initiate a resize but need finer control over the increase.