I have an e-commerce application based on Firebase backend, and I am using Firebase Cloud Functions to submit data to Elasticsearch for the sake of advanced search! Now that firebase reads costs a lot I am relying on ES(Elasticsearch) to read data such as products/users etc... So let's say that the reads operations are taking place on ES while write operations on Firebase.

How much this architecture is good? I mean is it good to rely on ES as primary database for read-only?

CodePudding user response：

I don't see any problems in this approach with proper Elasticsearch cluster configuration. Elasticsearch is very good for reading. I have projects that are using Elasticsearch as main read DB for years (more than 5 years). This is also e-commerce type projects.

However, for sure it depends on your load. If load is really big, Elasticsearch can also be expansive and may be will be better idea to add another type of fast cheap storage to handle reads (like DynamoDB in AWS, don't know exact GCP analog).