Elasticsearch should be used for its strengths as there are core principles based on which it is created. However, it does end up being used for other capabilities where it shouldn't be. When used in certain use cases, it will backfire, and those aspects are listed as disadvantages:
- Should not be used as a primary persistence store. So, apart from Elasticsearch, we need to have another store for actual persistence.
- Elasticsearch is not ACID-compliant out of the box.
- Even though Hadoop in a production environment is not run on commodity hardware, theoretically it does support this. In the production state, Elasticsearch inherently doesn't run on commodity hardware.
- Elasticsearch has Garbage Collection (GC) issues. ...
