Is It Recommended to Use Elasticsearch's API Directly When Camunda's APIs Are Insufficient?

I have a question about using Elasticsearch directly through its API, for example with the Elasticsearch Java Client (Using the Java API Client | Elasticsearch Java API Client [8.13] | Elastic).

Can I assume that the schema of the records stored in Elasticsearch by Zeebe will not change and thus will not break any code that relies on it? If not, what would be the right approach?

Yes, you can access data directly from elasticsearch. But be careful you mustn’t touch Zeebe’s indexes in elasticsearch.