Power Outage Recovery Checklist Generator

Generate a power-outage recovery checklist before the first outage

Automatic boot after a power failure fails in layers, so the reliable path is an ordered checklist that covers firmware recovery state, outage protection, service auto-start and a repeatable validation test — before the lights go out. This generator builds that list from your firmware options, UPS situation, service stack and database exposure, then hands you a copyable checklist ending in a three-cycle physical AC-restore test.

This tool is the pre-outage half of our power-recovery pair. If you have already configured State After G3 and want the full configuration walk-through with measured results, read the AC power recovery guide; for how every BIOS vendor labels the same option, see the AMI BIOS State After G3 guide.

Why the order is firmware → protection → auto-start → validation

Each layer assumes the previous one. A firmware set to S0 boot is useless if the PSU never sees AC again because a tripped UPS stays off. A machine that boots cleanly is still not recovered if the database was mid-write when power died. Docker containers without a restart policy stay stopped after the fastest boot. And a single successful power-on proves nothing — standby-power timing and Windows startup differ between cold runs, which is why the checklist ends with a three-cycle standard instead of a single test.

Keep the mechanism boundary clear: State After G3 fires on physical AC restoration, while Wake-on-LAN needs standby power and an armed NIC. The two fail for different reasons and are validated separately — the WoL checklist generator covers the network-wake path.

The three-cycle validation standard

Each cycle starts from a healthy state: remove AC input with a controlled switch, wait long enough for standby power to collapse, restore AC, then confirm automatic power-on without touching the case button. Time three milestones every run: boot, network up, and application health endpoint returning 200. We ran this on the EQi12 — the measured reference points this checklist cites are:

Commands each checklist step relies on

docker ps --format "table {{.Names}}\t{{.Status}}"   REM what actually restarted
docker inspect -f "{{.HostConfig.RestartPolicy.Name}}" <container>
Get-WinEvent -FilterHashtable @{LogName="System"; Id=41,1074,6008} -MaxEvents 10
Get-NetAdapter | Format-Table Name, Status, LinkSpeed

Event ID 41 (kernel-power) with no matching 1074 means Windows lost power uncleanly — exactly the case a UPS with a shutdown link exists to prevent. If your machine reboots into firmware setup instead of Windows after an outage, the State After G3 guide explains the BIOS-state behavior behind it.

Validation standard

The setup is done only when: three consecutive physical cycles pass with zero button presses, every cycle reaches a healthy application endpoint (HTTP 200 on our runs), the same milestones are timed each run, and the numbers are written down next to your own hardware baseline. Two of three is not a pass — document it as intermittent and look for a firmware or PSU cause before trusting it unattended.

Frequently asked questions