Low disk space watermarks

OpenSearch® relies on three indicators to identify and respond to low disk space.

Disk allocation low watermark

Defined by parameter cluster.routing.allocation.disk.watermark.low and the default value is set to 85% of the disk space.

When this limit is exceeded, OpenSearch starts avoiding allocating new shards to the server. On a single-server OpenSearch, this has no effect. On a multi-server cluster, amount of data is not always equally distributed, in which case this will help balancing disk usage between servers.

Disk allocation high watermark

Defined by parameter cluster.routing.allocation.disk.watermark.high and the default value is set to 90% of the disk space.

When this limit is exceeded, OpenSearch will actively try to move shards to other servers with more available disk space. On a single-server OpenSearch, this has no effect.

Disk allocation flood stage watermark

Defined by parameter cluster.routing.allocation.disk.watermark.flood_stage and the default value is set to 95% of the disk space.

When this limit is exceeded, OpenSearch will mark all indices hosted on the server exceeding this limit as read-only, allowing deletes (index.blocks.read_only_allow_delete).