[Node] Add retry on DA connection

[musitdev]

Summary

• RFCs: Link to RFC, Link to RFC, or $\emptyset$.
• Categories: any of protocol-units, networks, scripts, util, cicd, or misc.

Sometime during the docker container start, the full node start too early and the DA light node didn't have opened its entry point. It results in an error during the full node start, then a restart of the whole stack.
This scenario can repeat several times until the timing is correct.
This PR add a retry for this connection so that full node start is less sensible to timing issues during the start.

I've changed the error message when the config can't be parsed.

Changelog

• Ad a retry when the Http2 connection fails during init.
• add a better error message when the config can't be parsed

Testing

To test, you can start a leader and its follower:

CELESTIA_LOG_LEVEL=FATAL nix develop --extra-experimental-features nix-command --extra-experimental-features flakes --command bash  -c "just movement-full-node native build.setup.eth-local.celestia-local.test-followers --keep-tui"

The test execution should have no errors.

Outstanding issues

[0xmovses]

Thanks for the update on the error message, that is greatly needed.

[andygolay]

I ran the test and it passed; if the consensus is that the retry logic is needed then I'm willing to approve. Once concern is to have a "service available" condition in the orchestration to prevent this problem that causes the need for retry logic... maybe this could be something to add in a next iteration.

[mzabaluev]

Please fix the test workflow to not reference a nonexistent package.

[musitdev]

Fixed, there's some test that still fail like executor::execution::tests::test_execute_block_state_get_api but it's not related to this PR.

[mzabaluev]

Yes, I see the same failure on #1038, let's fix it separately.