/
Opensearch Disk Space Threshold (15% Free Minimum)

Opensearch Disk Space Threshold (15% Free Minimum)

Opensearch Disk Space Threshold (15% Free Minimum)

Pulse Opensearch requires a minimum amount of free disk space on its data drive to keep indexing data. If free disk space falls below the threshold, Opensearch stops writing to indices and Pulse will stop recording new monitoring data, change history and logs until the situation is resolved.

Minimum free disk space: Opensearch requires at least 15% free disk space on the drive hosting its data directory. Below this threshold, Opensearch will stop indexing and writing new data.

What happens when the threshold is breached

  • Opensearch blocks writes on all indices (read-only flood-stage protection).

  • Pulse Monitor and Pulse Application Server continue running, but new monitoring data, change history, and logs are no longer persisted to Opensearch.

  • Pulse Dashboards may appear empty or stale for the period the threshold was breached.

  • Errors of the form cluster_block_exception ... index has read-only-allow-delete block appear in the Opensearch logs and in pulse.log.

How to detect it

  1. Go to About → Status in the Pulse web client and check the Opensearch cluster status — it will typically show Red or Yellow with indexing errors.

  2. Check free space on the drive hosting the Opensearch data directory.

  3. Review pulse.log and the Opensearch logs for cluster_block_exception or flood_stage messages.

How to resolve it

  1. Free up disk space on the Opensearch data drive until at least 15% is available. Typical actions:

    • Remove old log files outside of the Pulse installation.

    • Cleanse old Pulse data — see Cleansing the Pulse data.

    • Reduce the frequency of Opensearch indices — see Changing the frequency of the Pulse Elasticsearch indices.

  2. Clear the read-only block on indices once disk space is available. Opensearch does not automatically lift the block — it must be cleared explicitly via an API call to remove index.blocks.read_only_allow_delete from affected indices.

  3. Restart the Pulse Opensearch service if indexing does not resume after the block is cleared.

How to prevent it

  • Monitor free disk space on the Opensearch data drive with a Pulse system-level alert (Administration → Alerts).

  • Set MaximumPulseDiskSpaceGB in Pulse.cfg to trigger an early warning before Opensearch's own threshold is hit.

  • Plan for long-term data retention by sizing the Opensearch drive based on expected index growth and review the data cleansing policy quarterly.

  • For environments with heavy logging, consider moving the Opensearch data directory to a dedicated drive.

Related pages