As part of the regular monitoring for customer db instances created in Neo4j Aura, we check to see when instances are approaching the maximum capacity of the physical storage associated with the db instance.
We have two alerting thresholds that we monitor on currently, when a db instance hits one of these two thresholds we raise a proactive ticket to notify customers a threshold has been breached and that some remedial action is required to prevent data corruption/loss and ensure the continuation of smooth operations continue.
In these cases it is important to note that deleting nodes / relationships will not reduce the disk footprint of your data, there are currently two courses of action to be taken,
- Perform compaction of your database store (an offline process).
- Resize the instance to a larger instance.
Perform Compaction of Database Store
Currently, Aura does not offer the compaction feature as standard and while we work to deliver this important management feature, the following knowledge base articles outline the procedures for performing a compaction of your database store.
- How to Estimate The Space a Compaction Can Save
- How to compact your database store
- Performing data store compaction (using an AWS EC2 instance)
Resize DB Instance to a Larger Size
If it is not possible to perform a compaction of the database store, the other option currently available to resolve the situation is to increase the size of your existing instance, as outlined in the knowledge base article Resize your Neo4j AuraDB instance.
For reference, please note that you can monitor storage usage within the Aura console as explained here.