ci: improve deploy workflows

[leovct]

Description

Find out why batches are not being verified

Erigon Setup

When deploying the CDK stack locally, we can see that the number of verified batches increases. The problem may come from the monitor script.

Looking at the CI job, it seems that the monitor script is stuck after the first request. There is a problem with the cast send.

It turns out the cast send cmd used to increase the number of batches is not handled properly by the new erigon node. We found a fix by reducing the gas limit used to 100000 instead of 643528 . It looks like this a regression from the legacy node.

ZkEVM setup

There were couple of things:

• A typo in the prover configuration and the same typo in the zkevm node library where we use zkevm-node as the sequencer type instead of zkevm.
• The zkevm sequence-sender is deployed twice.

Couple of quality of life improvements

• Speed up the deploy jobs by setting l1_seconds_per_slot to 1s instead of 12s.
• Add timeouts to each job, especially the deploy jobs - 20 minutes timeout.
• Transform the monitor-cdk-verified-batches local action into a script. Imo, it's easier to use and maintain.
• Improve the monitor-verified-batches.sh script.
  • It now requires an RPC URL instead of the name of the rpc service to query. This flag is mandatory.
  • It shows the script input values for debug purposes.
  • It shows the different batch counters (latest batch number on L2, virtual and verified batches).

References (if applicable)

https://0xpolygon.slack.com/archives/C060ZAVAU9K/p1722846740784669