ADA-C01 Question 362
Single answer4.4 Enable and manage logging and tracing.A Snowflake administrator needs to troubleshoot a Python stored procedure in a production database. The procedure intermittently fails, and the engineering team wants more diagnostic detail without changing the procedure code. The administrator must enable observability in a way that captures additional execution detail for the procedure while following Snowflake-supported logging and tracing controls. Which action should the administrator take?
- A
Set the TRACE_LEVEL parameter for the stored procedure to an appropriate value so trace events are captured during execution.
- B
Enable QUERY_TAG on the session running the procedure, because Snowflake uses query tags to generate procedure trace spans automatically.
- C
Increase the LOG_LEVEL parameter on the warehouse used by the procedure, because handler logging and tracing are controlled at the warehouse level.
- D
Create an event table and set AUTO_EVENT_LOGGING on the stored procedure so Snowflake automatically captures logging and tracing-related events for the handler.
Show answer and explanation
Correct answer: D
Explanation
Snowflake observability for procedures and functions relies on event tables as the storage destination for logs, traces, and metrics. Administrators can manage diagnostic capture using supported parameters, including AUTO_EVENT_LOGGING and TRACE_LEVEL, depending on the use case. In this scenario, the key requirement is to increase diagnostic detail for a Python stored procedure without changing the handler code. Creating or using an event table and enabling AUTO_EVENT_LOGGING on the stored procedure is the most appropriate administrative action because it allows Snowflake to automatically capture execution-related observability events. QUERY_TAG helps with query categorization, not handler tracing. Warehouse settings do not control handler logging/tracing in this way. Snowflake documentation on logging, tracing, and metrics for handlers, event tables, and object parameters such as AUTO_EVENT_LOGGING and TRACE_LEVEL supports this approach.
- A. Incorrect.
Incorrect. TRACE_LEVEL is a supported parameter for controlling trace event ingestion for procedures and functions, but by itself it does not enable automatic capture of handler logging and tracing detail when the requirement is to increase diagnostics without changing code. In Snowflake, automatic observability event capture for supported handler languages is associated with event tables and the AUTO_EVENT_LOGGING parameter. A candidate might choose this because TRACE_LEVEL sounds like the direct tracing control, but the scenario specifically asks for a supported way to capture additional execution detail without code changes, which points to automatic event logging.
- B. Incorrect.
Incorrect. QUERY_TAG is useful for labeling and later identifying queries in query history and related monitoring views, but it does not generate tracing spans or handler log events for stored procedures. This option reflects a common misconception that query metadata tagging is equivalent to runtime tracing. It can help correlate activity, but it is not the mechanism for enabling Snowflake logging/tracing observability.
- C. Incorrect.
Incorrect. LOG_LEVEL is not managed at the warehouse level for procedure handler observability. Warehouses provide compute resources, but Snowflake logging, tracing, and metrics collection for handlers are configured through supported observability objects and parameters such as event tables and object/session-level settings. This distractor targets the misunderstanding that all runtime diagnostics are tied to compute configuration.
- D. Correct.
Correct. To collect observability data in Snowflake, an event table must be available as the destination for logs, traces, and metrics. For supported handlers such as Python stored procedures, setting AUTO_EVENT_LOGGING on the procedure allows Snowflake to automatically emit logging/tracing-related events without modifying procedure code. This is the best fit for the requirement to capture more diagnostic detail in a supported, production-appropriate way.