What is MTTA?

The clock starts when the alert is dispatched and stops when a human acknowledges it in the paging system. It is the cleanest of the incident time metrics to measure because both endpoints are unambiguous events recorded by the paging tool.

MTTA is usually tracked separately for business hours and overnight, since the two look very different, and separately by severity, since a low severity notification acknowledged in an hour is not the same problem as a critical page acknowledged in an hour.

Acknowledgment time is dead time in every incident. Nothing is being investigated or fixed during it, and it is added directly to the total duration.

• Process health: a rising MTTA is an early indicator of alert fatigue.
• Paging reliability: long tails usually mean notifications are not reaching people.
• Rotation load: consistently slow overnight acknowledgment signals an unsustainable rotation.
• Escalation tuning: the distribution informs how long automatic escalation should wait.

Acknowledgment is easy to game, deliberately or not. A responder who acknowledges to stop the phone buzzing and then goes back to sleep produces an excellent MTTA and no response at all. The metric measures a button press, not engagement.

It also carries no information about outcome. A team with a two minute MTTA and a six hour MTTR has a fast paging system and a diagnosis problem, and MTTA alone would suggest everything is healthy.

MTTD measures how long it took to detect that something was wrong, running from the moment the failure began to the moment it was noticed. It measures monitoring quality.

MTTA starts where MTTD ends and measures human responsiveness once the alert exists. They fail for entirely different reasons: high MTTD means the monitoring did not catch it, high MTTA means the monitoring caught it and nobody responded. Fixing them requires different work, which is why they are worth separating rather than tracking a single time to response.