Verify recovery and close the job

Recovery is an observation over time

A successful restart is one event. Sustained recovery is normal behavior that you observe over a suitable period. How long that period is depends on the fault, the equipment and the approved procedure. A fault that comes and goes (an intermittent fault) may need a longer watch than a fault that appears right at startup.

Verification should return to the original complaint. If the complaint concerned accepted pool work, checking only the local interface leaves part of the problem unexamined. Compare the relevant measurements, time windows and operating state.

Resolution and explanation are separate conclusions

A machine can recover before anyone knows the root cause. A good service record can say both things: operation has been restored under the stated conditions, and the original cause is still unconfirmed. That tells the reader more than treating the recovery as proof of the diagnosis you favored.

Closing a job means recording what was verified, any remaining limitation and who owns follow-up. A recurring event should be connected to previous records so the next investigation can use the history instead of starting again.

Choose verification that matches the original failure

Turn the complaint into an acceptance statement: what must be true for the job to count as fixed. A missing-chain complaint needs evidence, after recovery, that every expected chain is listed and producing output. A missing pool worker needs a confirmed worker name and current evidence of accepted work. A handy green light cannot replace the measurement that first showed the problem.

Keep the time of the fix separate from the observation window. An hourly pool estimate can still include earlier downtime after work resumes. A recent normal reading may also come before the conditions that used to trigger the fault. Record which conditions you actually observed, what is still untested and who owns the next check.

For unresolved work, give the next specialist the asset identity, when the problem started, the saved evidence, any approved changes and the specific question you need answered. For any authorized parts swap or RMA (a return to the manufacturer or vendor for repair or replacement), follow the site procedures for tracking and handling parts.

Worked example: the green dashboard

After a part replacement, the local hashrate display looks normal again. But the assigned pool worker still shows no accepted work, and nobody has confirmed that the worker name belongs to this miner.

The local recovery is real evidence, but it does not fully answer the original complaint about production. Confirm the worker identity, then compare the local and pool readings. If the pool view is delayed, record that and assign the next check. Do not close the ticket just because one display is green.