ADA-C01 Question 363
Single answer4.4 Enable and manage logging and tracing.A Snowflake administrator is troubleshooting an intermittently failing Python stored procedure used by several ETL pipelines. The development team wants additional diagnostic detail, but the administrator must avoid generating excessive telemetry across the entire account. Which action is the MOST appropriate to enable targeted diagnostics for this procedure?
- A
Set the TRACE_LEVEL parameter on the specific stored procedure so trace events are captured only for that object during execution.
- B
Enable QUERY_TAG for the ETL role so Snowflake automatically records procedure trace spans and exception stack traces for all statements.
- C
Increase the LOG_LEVEL at the account level so all workloads emit more detailed logs, then filter on the procedure name afterward.
- D
Create an event table in the procedure's schema and Snowflake will automatically collect tracing data for that procedure without any additional configuration.
Show answer and explanation
Correct answer: A
Explanation
Snowflake observability for handler code and related execution diagnostics is controlled through event tables and configurable telemetry levels, including logging and tracing settings. Best practice is to scope increased diagnostics as narrowly as possible, at the object or session level when investigating a specific issue, rather than enabling verbose logging account-wide. TRACE_LEVEL is the parameter used to control trace event capture, while LOG_LEVEL controls log verbosity. QUERY_TAG helps identify queries but is not a substitute for tracing. Event tables serve as the storage destination for telemetry, but they do not by themselves enable trace generation. This aligns with Snowflake guidance on managing logging, tracing, and event table usage for targeted troubleshooting.
- A. Correct.
Correct. For targeted observability in Snowflake, tracing can be managed using object/session/account parameters such as TRACE_LEVEL. Applying the setting to the specific stored procedure scopes trace collection to that object rather than broadly increasing telemetry for unrelated workloads. This is the most appropriate choice when the requirement is to diagnose one intermittently failing procedure while minimizing account-wide noise and cost.
- B. Incorrect.
Incorrect. QUERY_TAG is useful for labeling and later identifying queries in history and monitoring views, but it does not enable tracing or automatically capture trace spans and exception stack traces. This option reflects a common misconception that query tagging is equivalent to observability configuration.
- C. Incorrect.
Incorrect. Raising LOG_LEVEL at the account level would be overly broad for the stated requirement and could increase telemetry volume for unrelated users, tasks, and applications. Although log levels are part of Snowflake's logging and eventing controls, changing them globally is not the best operational choice when only one procedure needs deeper diagnostics.
- D. Incorrect.
Incorrect. An event table is required to store log, trace, and metric data, but simply creating one does not automatically turn on tracing for a specific procedure. Additional configuration is needed, such as associating an active event table and setting appropriate logging or tracing levels at the desired scope.