[darkwater]

Yes, there is a long migration period (almost 2 years). Yes, if you are paying for support, they will help you a lot with the migration.

But if you are using a (costly) managed service it's because you don't want the hurdle of managing the service. Migrating off something is instead probably the most time consuming thing in a service lifecycle. So, even if it is normal to sunset products and people deal with it all the time, it's still a big PITA if you are affected.

[graemep]

> But if you are using a (costly) managed service it's because you don't want the hurdle of managing the service.

On the other hand part of the trade offs of using such services is that you have no control of when they will become unavailable. Its not just normal, its expected.

[Spivak]

I doubt very many businesses have anything but a surface level "vaguely gestures toward Azure maybe" plan for if EC2 ever shuts down.

[sokoloff]

I don’t have a plan for what happens if TCP goes away either, but that also seems fine.

EC2 and S3 are backbone/foundational enough that I expect they’ll be running the last year that AWS exists.

[graemep]

Most SMEs do not even seem to have backups with another provider.

I cannot see EC2 being shut down but other services might be, and there are all sorts of other issues. The idea that a cloud provider will just take care of everything is delusional. You still have work to do. It might be less work, but its there.

There was an Australian pension provider that recently had an issue with their main cloud account and services shut down. Luckily the regulator required they had backups with a different provider so they were able to restore everything. This is a business running something like AUD 125bn of other people's money.

https://www.theregister.com/2024/05/09/unisuper_google_cloud...

[arccy]

it's not just managed service where you can move to your self hosted version when they stop managing it, it's a fully proprietary stack that's just no longer available to use.