Acknowledgment Time SLA
< 5 min for sev 1.
Standard targets
Sev 1: under 5 minutes. Customer-impacting incidents need immediate response. Pager fires; on-call acknowledges within 5 minutes.
Sev 2: under 15 minutes. Significant but bounded. Time to acknowledge is forgiving; time to action is what matters.
Sev 3: under 1 hour during business hours; next business day off-hours. Issues worth knowing; not worth waking up for.
Measuring it
Time from page sent to acknowledgement received. Not from incident detected; that conflates detection and response.
P95 against target, not median. The pages that take 30 minutes to acknowledge are where things go wrong, not the typical 2-minute responses.
Per-engineer view: identifies people struggling. Per-rotation view: identifies systemic issues. Per-time-of-day view: identifies night-shift gaps.
How to consistently hit
Multi-channel paging. Phone call, app push, SMS, voice escalation. Each channel can fail independently; redundancy is the safety net.
Backup on-call with explicit escalation policy. If primary doesn't ack in N minutes, secondary gets paged. Tested monthly.
Test the path quarterly. Synthetic pages verify the full chain. Surprises during real incidents are expensive.
Reading the trends
MTTA trending up indicates degrading responsiveness. Causes: tool issues, on-call burnout, rotation understaffing.
MTTA trending down is usually positive. Watch for over-eager acks (acknowledging without engaging). Track time-to-action separately.
Per-time-of-day shows shift quality. Night shifts often higher MTTA than day; expected. Consistent night degradation suggests staffing gaps.
When MTTA is bad
First: paging tool reliability. Lost pages mean missed acknowledgements. The tool dashboard shows delivery rates.
Second: rotation health. Burnout, vacation gaps, off-hours coverage. Survey; staff up; address root causes.
Third: process. Routing accuracy. Severity classification quality. Often the MTTA target is reachable with better routing rather than more responsiveness.