Problem:

• For mongoDB Query analysis, it’s very important to track queries where a full scan of the dataset is happening.

  COLLSCAN is generally not a good approach for MongoDB because it involves scanning every document in a collection to find matching results. This can be extremely inefficient, especially for large collections or complex queries.

  Here are some reasons why COLLSCAN is inefficient:

  • High I/O Overhead: Reading every document from disk can significantly increase I/O operations, which can be a bottleneck for performance.

  • Long Execution Time: Scanning large collections can take a considerable amount of time, especially if the query is complex or involves many documents.

  • Scalability Issues: As the size of the collection grows, COLLSCAN becomes even more inefficient, making it difficult to scale the database effectively.

In PMM, We dont have option to filter queries that do COLLSCAN (full collection scan)or IXSCAN (index scan). Right now, only a very few options in PMM QAN for MongoDB.

The option to filter based on COLLSCAN / IXSCAN can be done via the option system.profile.planSummary field. For example:

Roadmap item

https://www.notion.so/percona/MongoDB-CollScan-tracking-in-QAN-129674d091f380dcb463c4178d29f695