Refs. sysadm-environment#5344

Related to sysadm-environment#5354

Use `terraform init -upgrade` to upgrade the provider

3 nodes are not enough. Given the requested resources
are not all specified and the nodes quite small,
gitlab eventually finishes to be unstable with
node killing pods to retreive some memory

Related to sysadm-environment#5177

Refs. sysadm-environment#5191

Ref. sysadm-environment#5176

This adds the resource group, one staging storage account and an associated
blob storage.

This prepares the requirements for the deposit migration to the dynamic
infra (in staging).  We are going to try and use an azure storage for the
tarballs the deposit receives.

Refs. sysadm-environment#5129

Related to sysadm-environment#4893

Related to sysadm-environment#4783

Related to infra/sysadm-environment#4695

An intermediary upgrade to 1.23.12 was made before upgrading to 1.24

Related to infra/sysadm-environment#4657

This properly stores all audit logs in the analytics workspace.

An intermediary upgrade to 1.23 was done

Related to infra/sysadm-environment#4657

Related to T4063

for example indexer-worker01 which has such configuration
```
disk_setup:
  ephemeral0:
    table_type: gpt
    layout:
      - [100, 82]
    overwrite: true
fs_setup:
  - device: ephemeral0.1
    filesystem: swap
mounts:
  - ["ephemeral0.1", "none", "swap", "sw,nofail,x-systemd.requires=cloud-init.service", "0", "0"]
cloud_config_modules:
  - [disk_setup, always]
  - [fs_setup, always]
```

Related to T4395

Related to T4395

Related to T4395

Related to T4395

The attach create_option of the disk changed from "empty"
to "attach" after the manual actions (unattach/reattach it).
The only way I found to avoid an error is to ignore this field.
Unfortunately, terraform doesn't not support ignoring a dynamic
resource's field, so it has to be pointed directly with the index

Related to T4222

It will allow to bootstrap the rancher manager without
manually installing a kubernetes on premise.

Related to T4046

Related to T4046