Post v9 ci chages

[faddat]

Background

A while ago I made a CI system for Gaia that would do statesync in CI using all of the supported databases.

Here is an example of it running:

https://github.com/tendermint/tm-db/actions/runs/2439480213

Suggested Design

I want to replicate this for osmosis v9 and begin to track rpc performance against live nodes in ci.

But this adds several jobs that will take over five minutes.

Acceptance Criteria

Mainly want to get thoughts on this from @ValarDragon , @p0mvn , and @catShaark since I know they're engaged on performance work.

[p0mvn]

I think this is a great idea. We already have WIP to enable e2e state sync tests in:

• refactor: e2e chain initialization to support spinning up full nodes in tests (backport #1587) #1613
• test: state-sync in e2e #1572

Since e2e is state synching from a local testnet, it should be pretty quick (under 1-2 minutes). There is going to be no support for several db backends in the beginning but it would be very useful to add it in the future. I think we should definitely enhance e2e with support for rocks or other database that we are recommending

[p0mvn]

Is there anything that we can't do in e2e and only with live nodes?

If yes and having live nodes is desired, I think we could enable this job to run only on merge to main branch to avoid letting developers wait for pull request CI to pass

[p0mvn]

Hi @faddat . I'm going to close this issue since it has been stale for over a month. Please let me know if you would like to continue discussing anything here