Improve the five availability components and their complexity in the reference architecture docs
The following discussion from !30781 (merged) should be addressed:
-
@meks started a discussion: (+8 comments) OK I think I found the reason why there is the lingering added focus on the trade-offs and the level of complexity. I came from &2833 (closed) then gitlab-com/www-gitlab-com#6817 (closed) then found this MR that Diana wrote last month !27046 (merged)
This MR was reviewed by Sid where we had a clear 4 level of availability complexity. In our current state we iterated so much that the original structure is not showing this increase step levels.
- Documentation: Better separation of scaling and HA concerns
- Separate scaling and high availability content
- Move reference architecture to scaling page
- Rename HA page to Availability
I am going to capture this structure in the most simplest form and delete the
Availability
page in the smallest iteration possible in this MR to prep for the meeting tmr per Eric's and Sid's update in #215161 (comment 334092532) -
@ljlane replied It would be great to have a conceptual guide to the components that leads in to the setup instructions. I have heard from customers that when following the docs they go through the setup instructions without really understanding the role of each component and why they need it. Not sure if we have covered that in existing MRs already, but I was thinking something like an overview based on function. For example, "Monitoring" section explains the components that monitor the health of master nodes in an HA cluster, "Failover" section that explains how failover works and the components that play a role in that. I can help pull this together. WDYT?