Skip to content

Act

Verification

Verification is the check that an action worked: MonitOps compares fresh telemetry with the action's recovery criteria after the run. A successful command is not a recovered service, so an incident resolves only when the criteria are met.

How it works#

Recovery criteria#

Each action in the catalogue declares what recovery looks like. A criterion can take one of these forms.

FormExample
A check passes over a windowThe synthetic checkout check passes on every run in 5 minutes
A signal is back within its expected rangeThe error rate of order-api is back within its baseline
A health stateA system reports healthy

A criterion names a signal in the catalogue (see Signal catalogue) and what must be true of it.

Fresh telemetry, not the command result#

After a run, MonitOps evaluates the criteria on telemetry that arrives after the run. The response of the command to the automation platform does not count: a call that succeeded can leave the fault in place, and a call that failed to return can have worked.

The outcome#

OutcomeResult
VerifiedEvery criterion is met. The incident resolves
Not confirmedAt least one criterion is not met. The incident stays open and says which one

Uncertain outcomes#

If the outcome of a run is uncertain, for example because the call did not return, MonitOps reconciles it before anything is retried. The action carries a stable operation identity, and the result is read back from the system. The action is never repeated on the assumption that the first attempt failed.

What resolves an incident#

An incident resolves when its alerts have cleared or, after an action, when recovery is verified. Alert recovery has its own rules: a numeric alert clears after consecutive evaluations meet the recovery threshold, and a check recovers after consecutive successful runs. Missing data never counts toward recovery.

What you see in the console#

The verification panel shows each criterion with its current reading and whether it is met. After the run for INC-20931, run RUN-5521 shows 20 / 20 checks passed over 5 minutes, the error rate of order-api back in baseline, and the incident Resolved, with the audit line for the approver.

Simulated data

Recovery verification

Recovery verified against fresh checks, not the command result.

20 / 20 passedSynthetic checkout probes over 5 min
order-api error rate back in baseline
INC-20931Resolved

14:09:52 ICT · run RUN-5521 · approved by Ploy K.

The run history shows the verification result for every run.

Limits and defaults#

  • Verification runs against the criteria the catalogue entry declares.
  • An unverified run leaves the incident open.

Failure states#

StateWhat happens
Criteria not metThe incident stays open and says why
Uncertain run outcomeReconciled before any retry

Last updated October 2026