Exams / SAP-C03 / Study guide / Operational automation and runbooks (EventBridge, Systems Manager, Service Quotas)
Operational automation and runbooks (EventBridge, Systems Manager, Service Quotas): SAP-C03 study notes and practice questions
What to know about Operational automation and runbooks (EventBridge, Systems Manager, Service Quotas) for the SAP-C03 exam, from the Operational Excellence and Automation domain (20% of the exam guide), then 1 free question to test it. Part of the SAP-C03 study guide.
SAP-C03 asks how to detect a problem, respond to it and keep the fleet patched across many accounts without people running scripts, and how to avoid hitting service quotas during growth or failover.
The event-driven response loop
Detect: CloudWatch alarms, AWS Config compliance changes, GuardDuty findings, AWS Health events (organization view in a delegated account) and service events all reach EventBridge.
Route: an EventBridge rule matches the event and sends it to a target.
Act: a Systems Manager Automation runbook, Lambda function or Step Functions workflow fixes the problem; SNS notifies people; an OpsCenter OpsItem records it.
Systems Manager for operations
Automation runbooks: AWS-provided (restart, resize, create AMI, enable encryption) or your own, with approval steps, rate control (concurrency and error threshold) and execution across multiple accounts and Regions from a central account.
Run Command for ad hoc commands on many instances; State Manager keeps configuration in place.
Patch Manager with patch baselines and organization-wide patch policies (set up through Quick Setup) scans and installs patches on a schedule; maintenance windows control when disruptive tasks run.
OpsCenter aggregates operational work items with related resources and suggested runbooks.
AWS Systems Manager Incident Manager is closed to new customers; new designs use alarms, EventBridge, SNS or chat notifications, OpsItems and runbooks.
Capacity planning and quotas
Service Quotas shows current quotas and usage; many quotas publish usage metrics to CloudWatch, so you can alarm before reaching a limit.
Quota request templates apply increases automatically to new accounts in an organization.
Raise quotas in the DR Region ahead of time; a failover that can’t launch instances fails.
Use On-Demand Capacity Reservations for capacity that must be there, and predictive scaling for daily patterns (see safe remediation and scaling).
How to choose
Scenario wording
Choose
“When disk usage passes 90%, expand the volume automatically”
CloudWatch alarm, EventBridge, Automation runbook
“Patch all instances in every account on a schedule with minimal setup”
Patch Manager patch policy through Quick Setup
“React to AWS scheduled maintenance events across the organization”
AWS Health organizational view with EventBridge rules
“Warn before we hit the vCPU limit”
Service Quotas usage metric with CloudWatch alarm
“Remediation must wait for an engineer’s approval”
Automation runbook with an approval step
“Run a fix on 500 instances without overwhelming them”
Automation or Run Command rate control
Common traps
Cron jobs on an EC2 instance that poll APIs and fix things.
Incident Manager for a new design; it is closed to new customers.
Requesting quota increases only in the primary Region.
Running remediation on every target at once with no error threshold.
Email-only alerts when the requirement is an automated response.
A telecom operator's provisioning service creates resources in one Region, and its account is close to the service quota for running EC2 vCPUs. The last two releases failed when quota was exhausted, and the operations team found out from the failed deployments.
The team wants an alert before the account reaches the quota, a documented response that starts without waiting for a person to notice, and a way to request a higher quota through the same automated path.
Which solution meets these requirements?
Choose one answer.
AA. Review the Service Quotas page each month, and open a support case by hand for any quota that is above 70 percent of its limit.
BB. Alarm on the Service Quotas usage metric for EC2 vCPUs in CloudWatch, and route the alarm through EventBridge to a Systems Manager Automation runbook that requests a quota increase.
CC. Request a large one-time increase of the vCPU quota in the Service Quotas console, and leave the quota there so that no alarm or automated response is needed for the rest of the year.
DD. Add a retry loop to the provisioning service that waits an hour after each quota error, so the release succeeds once capacity frees up.
Show the answer and explanation
Answer: B
A. Wrong: A monthly review is too slow to warn before a release fails, and the manual case is the step the team wants to automate. The option gives no automatic response.
B. Right: Correct. The usage metric and alarm warn before the quota is reached, and the EventBridge rule starts a Systems Manager runbook as the documented response. The runbook can file the quota increase request, so no one has to notice first.
C. Wrong: A one-time maximum increase may not match future use and does not alert the team if usage grows. The team also wants an automated response, which this option does not provide.
D. Wrong: A retry loop hides the shortage instead of warning about it, and the release can still miss its window. The option gives no alert and no documented response.
Explanation
A CloudWatch alarm on the quota usage metric warns early, and an EventBridge rule starts a Systems Manager runbook that files the increase request. Monthly reviews, a one-time increase and retry loops each leave the team unwarned or without a documented, automatic response.
Want more on Operational automation and runbooks (EventBridge, Systems Manager, Service Quotas)? The 5 full-length practice sets have 7 more Operational automation and runbooks (EventBridge, Systems Manager, Service Quotas) questions, 375 in all, none of them in the free exam. One payment of $14.99, no subscription. See the practice sets →
Study Operational automation and runbooks (EventBridge, Systems Manager, Service Quotas)