Question 290
At fictional Pinebrook Food Network, a Snowflake data and AI engineering team is conducting an accountable generative AI operations change using approved synthetic testing data instead of confidential customer records. The separately composed technical incident is: A nightly batch applies an LLM function to large batches of support messages and retries entire datasets after intermittent failures, generating duplicate calls and excessive consumption. Assume only these documented facts and use safe, authorized production-change procedures; exact current Snowflake feature syntax and region or account availability require current review. Which repeatable assurance or control is most relevant to preventing this situation?
Explanation
Writer-proposed answer: B (UNVERIFIED)
Preliminary individual option explanations — NOT expert validated
Correct Answer
Require per-record result status, retry caps, budget alerts and a restartable batch-processing runbook. LLM workload orchestration needs idempotency and fault isolation just as conventional data pipelines do.
Incorrect Answers
This response concerns Verified query repository for business metric questions, which is different from the stated failure. A nightly batch applies an LLM function to large batches of support messages and retries entire datasets after intermittent failures, generating duplicate calls and excessive consumption.
Cortex Search incremental refresh lag is not the cause of the reported failure: A nightly batch applies an LLM function to large batches of support messages and retries entire datasets after intermittent failures, generating duplicate calls and excessive consumption.
Sentiment interpretation with irony alone does not establish the needed evidence: Usage records and reconciled result counts show each eligible message is processed once and retries remain within the budget.