40 Elasticsearch interview questions with answers
40 Elasticsearch interview questions with detailed answers, sorted from basic to advanced. Free to practice, with progress saved automatically.
All Elasticsearch interview questions
What is an inverted index and why does it make full-text search fast?
What is the difference between shards and replicas, and how many shards should you choose?
How do `text` and `keyword` differ, and when do you use each?
What are the parts of an analyzer and why do you need to understand it?
How do `match` and `term` differ, and why does `term` often return nothing?
A document was just indexed but a search does not find it. Why?
Why is Elasticsearch usually not used as the primary database?
What is dynamic mapping and why is it dangerous in production?
How do query context and filter context differ?
What does the default relevance score (BM25) depend on?
Why is `from`/`size` capped at 10,000 and what replaces it?
Are `terms` aggregation results exact?
What goes wrong when an array of objects uses the default type?
What extra handling does Vietnamese search need compared with English?
A field type must change on a live production index. How do you do it?
How do you optimize loading a few million records into Elasticsearch?
What characterises Lucene segments and how do they affect operations?
What do index templates and ILM solve?
How do you make best-selling products rank higher in results?
How should autocomplete (type-ahead) be implemented?
Should search-term highlighting be done server-side or client-side?
Why is sorting on a `text` field rejected?
What are ingest pipelines for?
If you have replicas, do you still need snapshots?
Why must a browser never call Elasticsearch directly?
A query is slow but you do not know where. What tools do you use?
When do you use a runtime field?
Which index settings cannot be changed after creation?
How should data be synchronized from PostgreSQL to Elasticsearch?
How do you size the heap for an Elasticsearch node?
What does custom routing gain you and what does it risk?
How should indices be organized for a system serving thousands of customers?
How does vector (kNN) search differ from keyword search, and how should it be used?
What is the correct response to a `circuit_breaking_exception`?
How do you upgrade a live production cluster?
One node runs much hotter than the others. What causes it and how do you fix it?
What does Elasticsearch guarantee when a write request returns success?
What is the right way to export a few million documents?
How do you evaluate search quality after tuning ranking?
When should you not adopt Elasticsearch yet?
