SRE On-Call Practices & Runbooks 2 — Questions and Answers
Question 1: What is the primary purpose of an escalation policy in on-call rotations?
- To reduce the number of alerts fired
- To define who gets paged when the primary on-call does not respond (Correct answer)
- To automatically resolve incidents without human intervention
- To schedule maintenance windows
Correct answer: To define who gets paged when the primary on-call does not respond
An escalation policy ensures incidents are not dropped by automatically notifying backup responders when the primary on-call fails to acknowledge.
Question 2: Which metric best measures the quality of an on-call experience for engineers?
- Total number of services owned
- Mean time to resolve (MTTR) for all incidents
- Number of actionable alerts versus total pages received (Correct answer)
- Average team size
Correct answer: Number of actionable alerts versus total pages received
A high ratio of actionable to total pages indicates alert quality; noisy, non-actionable alerts degrade on-call quality and engineer wellbeing.
Question 3: A runbook step says 'restart the service if latency exceeds 2 seconds.' This is an example of what type of runbook content?
- Diagnostic step
- Remediation step (Correct answer)
- Escalation step
- Rollback step
Correct answer: Remediation step
A remediation step prescribes a corrective action to take once a problem condition is identified.
Question 4: What does 'runbook drift' refer to in SRE practice?
- Runbooks stored in multiple formats
- Runbooks that are rarely executed
- Runbooks that become outdated as the system changes (Correct answer)
- Runbooks written by multiple authors
Correct answer: Runbooks that become outdated as the system changes
Runbook drift occurs when documentation is not updated alongside system changes, causing instructions to become inaccurate or misleading.
Question 5: Which practice best reduces cognitive load for an on-call engineer during an incident?
- Granting the on-call engineer full production access
- Pre-writing runbooks with clear, numbered steps for known failure modes (Correct answer)
- Sending all logs to a single stream
- Disabling non-critical alerts permanently
Correct answer: Pre-writing runbooks with clear, numbered steps for known failure modes
Pre-written runbooks allow engineers to follow a structured response plan rather than improvising under stress, significantly reducing cognitive load.
Question 6: In a follow-the-sun on-call model, what is the main advantage?
- Engineers in one region handle all global incidents
- On-call responsibility transfers between regional teams to align with business hours (Correct answer)
- All engineers are on-call simultaneously
- Incidents are only addressed during peak traffic hours
Correct answer: On-call responsibility transfers between regional teams to align with business hours
Follow-the-sun rotations hand off on-call duty across time zones so engineers respond during their normal working hours, reducing night-time pages.
Question 7: What should happen immediately after an on-call engineer acknowledges a page?
- Close the incident and file a ticket for later
- Begin triage by assessing impact and consulting the relevant runbook (Correct answer)
- Escalate to management
- Disable the alerting rule that fired
Correct answer: Begin triage by assessing impact and consulting the relevant runbook
Acknowledging a page stops escalation timers, and the next step is triage — understanding impact scope and starting runbook-guided diagnosis.
What is the primary purpose of an escalation policy in on-call rotations?