You can include a retention policy in an SLM policy to automatically delete old snapshots. Retention runs as a cluster-level task and is not associated with a particular policy’s schedule. The retention criteria are evaluated as part of the retention task, not when the policy executes. For the retention task to automatically delete snapshots, you need to include a retention object in your SLM policy.

To control when the retention task runs, configure slm.retention_schedule in the cluster settings. You can define the schedule as a periodic or absolute cron schedule. The slm.retention_duration setting limits how long SLM should spend deleting old snapshots.

You can update the schedule and duration dynamically with the update settings API. You can run the retention task manually with the execute retention API.

The retention task only considers snapshots initiated through SLM policies, either according to the policy schedule or through the execute lifecycle API. Manual snapshots are ignored and don’t count toward the retention limits.

To retrieve information about the snapshot retention task history, use the get stats API:

GET /_slm/stats

The response includes the following statistics:

{
  "retention_runs": 13, 
  "retention_failed": 0, 
  "retention_timed_out": 0, 
  "retention_deletion_time": "1.4s", 
  "retention_deletion_time_millis": 1404,
  "policy_stats": [
    {
      "policy": "daily-snapshots",
      "snapshots_taken": 1,
      "snapshots_failed": 1,
      "snapshots_deleted": 0, 
      "snapshot_deletion_failures": 0 
    }
  ],
  "total_snapshots_taken": 1,
  "total_snapshots_failed": 1,
  "total_snapshots_deleted": 0, 
  "total_snapshot_deletion_failures": 0 
}