Today I was replying to one of the tasks regarding DigitalOcean accounts, and I was wondering: why do we have that many DigitalOcean accounts?
When we created these accounts, we were focusing on separating the accounts based on their environment and role (like internal infrastructure prod/staging, grove prod/staging, etc.) We have 5 accounts at the moment, but I feel it is way more than necessary.
Although we can name accounts and switch between them, I would like to avoid the situation that we had with OVH, i.e., splitting resources across ~9 accounts.
I propose the following:
Have an OpenCraft infrastructure account (having both staging and production resources there)
Have a Hosting account (currently called Grove) and have both staging and production resources hosted there. Also, Listaflow resources would end up here too.
[CLIENT NAME] cluster(s) where client-specific resources would go. Currently, we only have of these accounts for the hosted sandboxes.
Having staging and production resources in the same account poses no security concerns, and we could have a leaner setup.
@team as a result of this conversation, I’m going to shut down the OpenCraft staging infrastructure (today) as it serves no purpose that couldn’t be done on the production cluster itself. If we need to do any risky changes, we can setup a test cluster beforehand anyway.
The Grove Staging account and cluster will not be touched until the end of sprint 350. So this change only affects the internal infra we have (and not even that as staging is not used).
@jill This change for now is meant to be only for the internal infrastructure cluster we have. “The Grove Staging account and cluster will not be touched” now. Also, we will make sure the Tagging instance won’t be negatively affected (i.e., we migrate it beforehand or not doing the migration at all until the service is needed).
to reducing those ;p