Query Logging Strategy¶
DeltaTraceDB itself does not handle query logging automatically. Instead, it is the responsibility of the server-side application to determine what information should be recorded for traceability, auditing, or debugging.
Because each query is represented as a JSON-serializable object, logging can be implemented simply by saving the received query JSON before executing it.
Typical logging strategies include:
Full Logging: Save every query exactly as received. Useful for debugging and complete historical traceability.
Filtered Logging: Record only certain types of operations (e.g., writes only). Useful for reducing storage size.
Summarized Logging: Store metadata (timestamp, user ID, query type) but not the full query payload. Useful for production systems where performance and privacy matter.
Regardless of the chosen strategy, we recommend the following:
Always store write operations (e.g.,
add,update,delete) so the entire database can be reconstructed.Optionally store read operations depending on privacy policy and observability needs.
Since queries are just JSON objects, persistent logs can be written to files, cloud storage, or any logging backend with no schema conversion required.