A logistics job already pages operators through an SNS topic when a queue-depth alarm fires. The job also needs to page that same topic when the nightly run finishes cleanly. What should the job do?
Select an answer to reveal the explanation.
Short Explanation
The paging channel is already built, so you do not invent a fake fire to ring the bell. The job should publish its own completion message to the topic the alarm uses. A second alarm on a log word is a brittle substitute, and CloudTrail is not where an application announces that it finished.
Full Explanation
An SNS topic can accept publishes from an alarm and from the application. The nightly job already knows when it finishes, so it should call publish on the same topic. Existing subscriptions then deliver that message through the channel operators already watch. A second alarm that searches a log line for a word adds a delay, a parsing rule, and a second thing to keep correct, when the job can state the fact directly. Creating a fresh email subscription inside the job bypasses the topic the on-call rotation already manages, so the next rotation change misses this job. CloudTrail records AWS API calls. An application completion is not a management event, and writing a log line there is not how subscribers are paged. Caveat: the job's role needs permission to publish to that topic, and the message should be distinct from an alarm payload so operators can tell a clean finish from a page. Operational check: run the job in a test account, confirm one publish arrives on the existing subscription, and confirm an alarm publish to the same topic still arrives.