Reframe and complete GitLab HA (Scaled And High Availability) training
This issue is here to track improvements and changes to be made to the Scaled and High Availability training.
The problem
In its current state, the HA training is unfinished and generally unachievable by an individual support engineer. This is because the second step of the training is to spin up a minimal HA architecture which according to our documentation, requires a minimum of 30 nodes to truly be a true highly available. This is impossible to do on a local machine and overly expensive on cloud platforms.
Proposal
The training needs to be reframed to focus on setting up a distributed GitLab components and linking together to create a functional instance, with a theorhetical module on GitLab HA. Something like the following:
- Stage 0: Create and commit to the module
- Stage 1: Learn about GitLab Reference Architectures
- Stage 2: Spin up a 1,000 user reference architecture
- Stage 3: Learn about High Availability concepts
- Stage 4: Learn about GitLab HA architectures
This issue will track the MRs and discussion around the re-write of the module.
Helpful Links
https://docs.gitlab.com/ee/development/scalability.html#reference-architecture-overview https://docs.gitlab.com/ee/administration/reference_architectures/