Network cutover readiness checklist

What to check before you change a live network or system. Most cutovers that go wrong were missing something simple: a rollback trigger nobody agreed, a site key nobody had, a test that left no evidence. This checklist covers the six areas we review before every cutover.

Download the checklist (PDF) Get an independent review

Why a checklist still matters

Experienced teams don't fail on the hard parts. They fail on the small things everyone assumed someone else had done. A short, shared checklist turns those assumptions into named owners and written answers, a day before the change instead of at 2 am.

Use it in your readiness meeting, a day or two before the window. Anything you can't tick is either fixed, accepted as a risk by the person who owns the go / no-go call, or a reason to move the date.

The checklist

1. Plan and method of procedure

Every step has an owner, a planned start time and an expected duration.
Dependencies and hold points are marked, with the person who releases each hold point.
The change window fits the plan, with time kept back for a rollback.
The pre-change state is captured: configurations backed up, baseline readings recorded.
Vendor and third-party steps are confirmed in writing, with names and phone numbers.

2. Testing

Pre-checks are written down and run the day before, and again on the day.
Each test has a pass criterion, not just "check it works".
Every test leaves evidence: a reading, a screenshot, a log or a signed record.
Customer-facing services are tested end to end, not just the links in between.
Someone who did not make the change confirms the acceptance tests.

3. Rollback

Clear trigger points: which result, or which time on the clock, means you roll back.
Rollback steps are written out in full, not "reverse the steps above".
Time to restore is estimated, and it fits inside the change window.
The rollback has been rehearsed or tested where that is practical.
The person who calls the rollback is named, and so is their backup.

4. Communications

Customers and users know what will happen, when, and what they will notice.
Operations, the NOC and on-call teams have the plan and the contact list.
One bridge or chat channel is set up for the night, run by one person.
Update times are agreed: for example every 30 minutes and at each hold point.
Messages are drafted in advance for success, for a rollback and for an overrun.

5. Sign-offs and authority

The change is approved through your change process (CAB or equivalent).
The go / no-go decision owner is named, and the criteria are written down.
Contract and client requirements are checked: notice periods, approvals and outage limits.
Everyone knows who can stop the work on safety grounds, and that they will be backed.

6. People and logistics

Site access, keys, inductions and permits are arranged for every site involved.
Spares, tools, test gear, cables and charged laptops are checked and on site.
There is a shift plan with breaks, so nobody runs the critical steps exhausted.
Risk assessments or SWMS are in place for the on-site work.
The contact list, with mobile numbers, is checked the day before.

Five red flags we see most often

"Rollback: reverse the above." If it isn't written step by step, it hasn't been thought through.
No time to restore. Without it you cannot know whether the window is big enough to fail safely.
Nobody owns the go / no-go call. On the night, the decision drifts to whoever is loudest or most tired.
Tests with no evidence. "Looked fine" is not an acceptance record, and it will not help the next shift.
The same people on shift for 14 hours or more. The riskiest steps often come last, when people are most tired.

Running the go / no-go call

Status of every pre-check, read out against the plan, not from memory.
Open risks and issues, each with an owner and a yes or no: do we proceed with it?
Rollback confirmed: triggers, steps, time to restore and who calls it.
People and sites ready: access, spares, shifts and contacts.
The decision, made by the named owner and recorded with the time.

After the cutover

Keep watch. Agree a hypercare period with extra monitoring and a named contact.
Close it properly. Record the evidence, update the as-built records, and close the change.
Learn from it. A short lessons-learned session while it's fresh, written down for the next one.

Want a second pair of eyes before go-live?

Our Cutover Readiness Review checks your plan, testing, rollback, communications and sign-offs, and gives you a clear go / no-go view and what to fix first. Fixed scope, agreed up front.

Book a Cutover Readiness Review What the review covers