Amazon Elasticsearch allows for the easy management of an Elasticsearch cluster. One important Elasticsearch metric monitored by Blue Matador is the amount of disk space available in the cluster.
If the total amount of used storage in the cluster is too high, Blue Matador will create an event. Running out of disk space in a cluster can prevent shards from replicating and at critical levels the cluster will block all writes to a node that is out of disk. It is important to get ahead of disk space issues before they become critical to prevent this kind of outage.
One way to resolve disk space issues is to increase the amount of disk space available. This can be done by configuring the volume size or adding more nodes to allow an index’s shards to be stored on the new nodes. If your index is large and you are on a newer version of Elasticsearch (6.2+) then you can use the split operation to create a new index with the same data and more shards.
Another option is to clear up some used disk space. Deleting old indexes, merging segments or optimizing your index mappings can all significantly reduce the amount of disk and memory used in Elasticsearch. Which approach you take is entirely dependent on your data and your usage of Elasticsearch. The
GET /_cat/allocation?v API call will help you determine how much space each index uses.