Bump MySQL version to 8.0

[agnivade]

https://mattermost.atlassian.net/browse/MM-55589

[agnivade]

Strangely, this still creates a 5.7 instance. Need to investigate some more.

[agnivade]

Woops my bad, I changed the param group, not the engine version. 🤦 Monday morning.

[streamer45]

Thanks 👍

[agnivade]

@agarciamontoro - Wondering if @DHaussermann should give this a quick smoke test just to ensure there is no regression?

[agarciamontoro]

Woohoo! Thanks, Agniva 🎉 And yes, we should test this similarly to how we tested the Postgres bump in #590

[agarciamontoro]

Oh, huh, the one for postgres should have been aurora-postgresql14 🤔

[agnivade]

Closing for now. To be reopened on Feb 2024.

[agnivade]

Reopening this now! FYI @agarciamontoro @streamer45

[agarciamontoro]

@agnivade, next step here would be for @DHaussermann to test it, or is there anything else we need to do before that?

[agnivade]

Yep, nothing else needed.

[DHaussermann]

@agnivade everything seem normal with the Load testing process.

• The deployments have changes as expected.
  
• The deployed instance seem to function normally and I could see the regular activity from the load test as expected
  
• The data generated in Graphana for both bounded and unbounded seems typical
  

No issues found LGTM!

I couldn't think of a simple way to benchmark the MySql versions against each other. As this is configured at the Terraform level, both base and new will always use the same version of MySql. @agarciamontoro, am I mistaken? I feel like I did this for Postgres once a few months ago but, can't recall how.
Also, I understand that comparing performance was not the goal here but I was curious on if there are appreciable performance improvements.

[agnivade]

Thanks Dylan!

[agarciamontoro]

I couldn't think of a simple way to benchmark the MySql versions against each other. As this is configured at the Terraform level, both base and new will always use the same version of MySql. @agarciamontoro, am I mistaken? I feel like I did this for Postgres once a few months ago but, can't recall how.

A late answer, sorry, but here it goes: the only way is to run both tests independently and then manually generate a report for each of them (with report generate) and compare them (with report compare). It's what the comparison sub-command makes automatically, but you'd need to do it manually.