Add the tiers of your escalation chain — who is notified, how, and how long to wait for an acknowledgement. See when each person hears about an unanswered alert and export the policy.
Each row is a tier: who is notified, the channel, and the minutes to wait for an acknowledgement before moving to the next tier.
Runs entirely in your browser — nothing you enter here is sent to a server.
An escalation policy is the answer to "what happens if the person on call does not respond?". It is an ordered list of tiers; each tier names who is notified, by which channel, and how long the system waits for an acknowledgement before it moves on. Without one, a missed page at night means an outage that lasts until morning.
The policy is a trade-off between two failures. Escalate too slowly and an incident sits unanswered. Escalate too quickly and the second person is woken before the first has had time to pick up the phone — which teaches everyone that pages are someone else's problem.
Begin with a push notification, then a phone call to the same person a few minutes later. Most alerts are acknowledged at the first step.
Enough time to wake up, find a laptop and acknowledge. For the most critical services keep the first tier to five minutes.
The last tier is a manager or a team-wide call. Its purpose is to make sure a human knows, not to fix the problem.
A critical alert deserves the full chain. A low-severity one should create a ticket and never phone anyone.
An escalation policy that has never fired is a guess. Run a drill each quarter: trigger a test alert, let the first tier ignore it and check that every later tier is reached on time and on the right channel. Phone numbers change, people leave, notification settings get muted — and you only find out when it matters, unless you test.
A set of rules that says who is notified about an alert, in what order and after how long, when earlier responders have not acknowledged it.
Five to fifteen minutes per tier is typical. Shorter for services where every minute of downtime is costly, longer where a slower response is acceptable.
Three is a sound minimum: the primary on-call, a secondary, and a manager or the whole team. Splitting the first tier into "notify" and "call" steps for the same person is a common refinement.
A rotation decides who is on call at a given time. An escalation policy decides what happens with an alert: it usually refers to rotations in its tiers, such as "primary on-call" and "secondary on-call".
Upzilla checks your sites, servers and APIs around the clock and alerts you the moment something breaks. 10 monitors free, no card required.