The deployment of a cloud migration – or the ‘go-live’ – is a nerve-wracking time. For me, go-lives have often meant hundreds of hours of work, many late at night. Even just thinking about a go-live quickens my pulse. Hopefully these articles are helping you to manage that feeling a little. 

So, this time, I’m going to offer some advice to help you prepare for a go-live. Hopefully this will also ensure it’s a successful one.

It should be a ‘all hands on deck’ situation

Make sure you plan for this, because it’s the reality. Don’t go it alone – everyone should be together. Even if some teams think otherwise, they still still at least be on-call. This is especially true in your first few migrations when it’s really hard to know exactly what could go wrong. Keep everyone involved!

Plan as far ahead as possible

Nobody wants to lose a night or, even worse, a weekend, to a work activity. By planning ahead, you’re not only reducing the risk of this happening, but people will be less upset if they do have to work out of hours – simply because they’re somewhat prepared to do so.

Understand what ‘complete’ looks like

When is a cloud migration ‘complete’? This is an important factor to consider and agree upon with stakeholders. Once everyone understands exactly what ‘complete’ looks like, you’ll then know when you can leave (or log off from) work that night. So will everyone else. Also, if people want to leave early, make sure they know what they have ownership of. This way, if something goes wrong, you know who to call.

Understand what a ‘rollback’ really is

Let’s be honest, a rollback is never easy – so don’t take it lightly! It’s important to understand what a rollback entails in your situation, and by what criteria it should be executed. So, somewhere in your migration plans you should have a ‘point of no return’. This is the point you have to move forward from regardless of what happens. In other words, a rollback cannot happen from this point onwards.

Utilize a clone environment

Walking through the steps of the cloud migration in a clone environment should be a part of testing anyway, but can also be a simple walk-through before the cutover point. Do this as much as necessary.

Don’t forget post-migration checks!

Even many days after a cloud migration, it’s important to check that everything’s still operating as it should be. You don’t want it to be weeks before you realize something’s wrong (e.g a process that hasn’t been running since the migration). Or, even worse, a cloud bill that turns out to be double what was expected. The earlier you catch these, the better.

Remember: the cutover can be really scary and cause a lot of anxiety. If you prepare well, practice, and communicate efficiently with other teams, you should do just fine.

Let me know what you think of my advice down in the comments below, I’d love to hear from you. Feel free to also share advice and experiences of your own.