All lessons Leer en español

Security in depth · Unit 29 · Lesson 8 of 9

Define what makes recovery complete

Decide whether a reachable service has met its agreed return-to-operation checks.

4 minreadyShort lesson

Helpful before thisIncident response: decisions under uncertainty

See all lessons in this topic

After this lesson you can

  • Separate availability, integrity, permission, and monitoring evidence before declaring recovery complete.

A page loading successfully is one recovery signal among several.

Agree what complete means

Recovery exit criteria are observable conditions required before declaring a defined recovery stage complete. They should cover necessary service behavior, data integrity, access, and monitoring. A time objective helps prioritize work; it does not turn an unmet security or integrity requirement into a pass.

Assume a fictional booking service must restore all six hundred records committed through 11:00, preserve read-only support access, and provide telemetry no more than two minutes old before normal operation. No deviation from these criteria has been approved.

Restored service → Agreed exit criteria → Monitored return to operationRestored serviceAgreed exit criteriaMonitored return tooperation
Restored availability must be compared with all agreed criteria before a monitored return to normal operation.

Supplied restoration results

  • R1: The application loads and the approved booking workflow functions in the recovery check.
  • R2: Reconciliation finds 592 of the 600 required records through 11:00. The eight missing records are inside the required recovery scope.
  • R3: The support role reads its permitted data and its attempted change is correctly denied.
  • R4: The latest usable telemetry is ten minutes old; no newer delivery is established.

Preserve the failed checks

R1 and R3 support their stated criteria. R2 is an integrity gap; it cannot be dismissed as acceptable loss outside the recovery window. R4 fails the freshness requirement, so a quiet dashboard cannot establish healthy post-recovery observation.

Report partial restoration and assign reconciliation and telemetry follow-up. A separately authorized degraded service could have different conditions, but it would need its own explicit decision and limits; it is not the same as declaring normal operation.

The final recovery record should include the verification results, approving owner, residual limitations, and ongoing observation plan.

Model decision: Keep normal-operation approval pending until the missing records and fresh telemetry are verified.

Terms you met

Recovery exit criteria

Check yourself

No timer. No penalties. Read the explanation and try again whenever you like.

  1. What recovery status is supported?

    Show the answer

    Correct answer: Partial restoration; the integrity and telemetry criteria prevent declaring normal operation. The page works and the role check passes, but the agreed data count and freshness requirements remain unmet.

Try it

  • WriteWrite a four-row recovery status for R1-R4 and a final decision. Mark availability and role checks passed; integrity failed and telemetry freshness unmet. Name the evidence still needed for normal operation.
References