Allow to manually enter S3 bucket names

[agowa]

While going through the first run setup it should allow to manually enter the bucket names as well as to specify a separate API Key and Secret. S3 providers like Backblaze allow to restrict the bucket listing API calls, so that listing all available buckets is not possible and throws an error. Compatible S3 providers sometimes also just allow for one API Key + Secret pair to access a single bucket, therefore a separate key is needed per bucket. Also from a security perspective, I don't want to have a single user to access more as needed. I feel way more comfortable if the credentials for one application can only access the data that that application needs as well as when (assuming breach/compromise) those credentials don't even know what else is accessible within the same account/tenant.

[vmatekole]

S3 services usually require globally unique bucket names. This automatic naming at setup can fail silently when a bucket already exists. Generally the S3 setup is sub-optimal and is prone to failure and can only be fixed by manually changing the config or reinstalling cells.

[cdujeu]

We cannot handle every cases via the UX. You can simply use the YAML/JSON configuration to set the buckets you want.