feat: add generic s3 backup options

[pat-s]

So far this was only able when using cloud-provider. Now, one can set backup_provider to one of the mentioned options and backup to S3 through pgbackrest when the autobase cluster has been deployed via Ansible.

Feel free to tweak to your liking, I just had this already in use since some time and wanted to part the idea upstream.

In addition to setting backup_provider, this also requires setting the respective access keys, storage name and other vars required to configure the S3 backend.

[pat-s]

relates to #916 and is a subset of #919

[vitabaks]

In fact, we’re enabling auto_conf here even when the cluster is not fully deployed in the cloud via autobase. I’m not entirely sure why this is necessary without functionality like automatic S3 bucket creation — that was the original idea behind using auto_conf.

If the user creates the bucket manually, they can also define the configuration themselves based on the example in the documentation. In this case, it actually makes more sense to do it manually, since the bucket name may differ from the default suggested in auto_conf tasks

I thought it would be better to finish #919, where we make it possible to deploy bucket in the right cloud (regardless of the cluster deployment method) and automatically configure the backup configuration there..

[pat-s]

There are many ways to approach this.
This was just one which I used to quickly get it going on my end. Feel free to approach it differently, just wanted to share this one, so others could possibly adapt it.

[pat-s]

@vitabaks Shall we close here? From what I understood, you don't want this approach to be added, right?