Technical discovery on how to improve CASEI build time

[heidimok]

Is your feature request related to a problem? Please describe.
We ought to improve build time of the front end as part of hand off because it's slow

Describe the solution you'd like
Look into how we can speed up front end build

Things loosely considered

• Updating Gatsby to open up the possibility of incremental builds. It might not be the only thing we can do to improve build time but it's a good guess for a start.
• Can also think about alternatives that might give us faster build times.

What is done?

We want to time box the discovery aspect so that it doesn't take more than 1-2 days.

Depending on the solution, we can close this issue by:

1. Implementing a solution for improved build time
2. Proposing a solution for improved build time (and creating new issue(s) that are specific to that solution)

[naomatheus]

I'm 90% that I've been able to resolve dependency issues. (Documented in this PR).
However, I'm getting timeouts on the source and transform nodes steps of Gatsby build. I'm fairly certain it's not a connection issue, but instead something underlying.

cc @Tammo-Feldmann

[naomatheus]

Going to move notes to the PR above ^

[naomatheus]

Update:

Progress has been logged in PR #562. The PR will be rolled back to commit [4edcddd](https://github.com/NASA-IMPACT/admg-casei/pull/562/commits/4edcdddad68e7bea89a2403155c0fd02a10b4671). Gatsby v5 upgrade is complete at this point in commit history. This brings CASEI's codebase up to date with current standards and should improve overall stability of production deployments and Continuous Integration.

Other PRs and comments to this PR were discussing implementing the Slices API for Gatsby. Implementing Slices allows the CASEI site to be "split up" into discreet components that will bring down the build time by building CASEI in "Slices", as opposed to on a per-page build process.

Caveat and moving forward

Implementing Slices does require non-significant refactoring so I believe this should be split into a separate issue, such as "Implement Slices API in CASEI" or something of that nature. I have discussed this with @Tammo-Feldmann who may be the most knowledgeable engineer for the slices implementation task and I would like to support.

[heidimok]

@naomatheus Thanks, so I think we can consider this issue done.

However, would you be able to create that separate issue to "Implement Slices API in CASEI" as mentioned above for the backlog on and then close out this issue?

[naomatheus]

Updates

This issue will closed with the following result:

• One new issue/ticket created for implementation of the Slices API that will be added to Backlog
• Improving CASEI build type has two components
  • Upgrading CASEI from Gatsby v4 to Gatsby v5.
  • Implementing the Slices API to "chunk" components at build time and increase overall build time

Result 1:

• Upgrade from Gatsby v4 to v5 has been split out into this PR (Issue 524/isolate gatsbyv5 #568)

Result 2:

• New issue created for implementing the Slices API in CASEI. For this, some material is left over in this PR's commit history. That being said, starting from scratch after v5 is merged to develop is also reasonable.