DOP-C02 Practice Question 90: Run Command Rate Control
AWS Certified DevOps Engineer – Professional (DOP-C02) · Test 9 · question 10 of 10
A DevOps engineer needs to send a Systems Manager Run Command document to 2,000 managed instances to restart an agent process. To avoid overwhelming the fleet and to stop the rollout automatically if the change starts causing widespread failures, the engineer wants to limit how many instances run the command at the same time and have the command stop being sent to additional instances after a defined number of failures. Which Run Command settings should the engineer use?
Show answer and explanation
Answer:
Run Command exposes rate controls directly: max concurrency limits how many managed nodes run the command at the same time (as an absolute count or a percentage), and max errors stops sending the command to further nodes once that many invocations have failed, which together match the requirement without any custom tooling.
A is wrong because those settings belong to Automation runbooks, not to a plain Run Command invocation, and Run Command does have its own native rate controls despite the option's claim otherwise. C is wrong because disabling the service role is a blunt, disruptive action that would also break other legitimate Systems Manager operations, and it is not how rate control is implemented. D is wrong because there is no such IAM condition key for limiting concurrent targets per SendCommand call; concurrency and error thresholds are configured as rate control parameters of the command itself, not enforced through IAM policy.
Reference: Run commands at scale