[cadamsdotcom]

One amateur moment doesn’t make a service’s management amateur.

If you run it yourself there’s a chance you will trade the mistakes made by DO for different mistakes made by your own team - and still have similar overall reliability.

[Nextgrid]

Simply moving the time at which the mistakes occur can be extremely valuable. Doing it yourself means you can say "no touching the server during business hours". You can't guarantee that with a provider.

Even if you work out that you cannot do better, at least you are no longer paying the insane premium of the managed highly-available service (since it's not actually capable of delivering).