The test was occasionally failing because of a slow test.

ctb: Fix devnet

Increase sequencer window & sequncer drift

The ss was not previously being shut down cleanly, which
can result in a corrupted db state. This makes the shutdown
more clean.

op-chain-ops: allow config of LevelDB in `op-migrate`

This should hopefully help some of the reorgs we've seen on the devnets.

Two new flags are added

- `db-cache`
- `db-handles`

These are passed through the the LevelDB constructor

op-migrate: various sanity checks + cleanup

Adds cleanup to the migration script, ensures that
config is sane.

We cannot allow for non monotonic block time. We also
try to make eip 1159 stable at start and use a hardcoded
initial base fee.

Add additional logging

op-chain-ops: check for transition block config

Adds checks for transition block config. We do not
want these values to be 0 because it will mess with
eip 1559 in the beginning.

op-chain-ops: add no check to eth migration

op-chain-ops: migrate db monotonic timestamp

Adds support for finalizing legacy withdrawals after the Bedrock
migration. Cleans up some SDK code at the same time.

Ensure that the timestamps are monotonic when
creating the bedrock transition block

op-chain-ops: add 1559 config checks

For both the denominator and elasticity. They cannot be 0.

op-chain-ops: base fee check

op-chain-ops: safety with params

Check that the L2 genesis block base fee per gas
is not set to `nil`. If it is, it will cause a segfault
in `op-geth` on startup after the migration.

It is configured with the key: `"l2GenesisBlockBaseFeePerGas"`

Check for `nil` when getting the chain params.
Also log info about the chain params

Without it, there will be a `nil` dereference.
Each network should be added explicitly

op-migrate: use correct ancient path

op-chain-ops: better error handling in op-migrate

Remove changesets dir from codeowners

Prevent `nil` dereference by explicitly checking for `nil`.
Add in logs to give better understanding of what the script is
doing.

Merge pull request #4126 from ethereum-optimism/11-30-op-chain-ops_Migration_should_not_use_readonly

op-chain-ops: Migration should not use readonly

maint(ctb): clean up deploy config

op-chain-ops: more cleanup in migrate script

The `op-migrate` script had some leftover stuff
in it from early iterations on its development.
The leftover stuff was cleaned up to improve the
overall UX of the script. Now code should not need
to be commented out + rebuild the binary, it will
automatically detect when the script has already
ran on a datadir.

Also pass through the no check arg so that we can
re-enable the checking and dynamically change if
we want to run the data validation as part of the
migration.

op-chain-ops: set initial message passer nonce

The `L2ToL1MessagePasser` nonce is set to `type(uint32).max`
initially to prevent nonce collisions with legacy withdrawals
that have been upgraded. The max value for the nonce is
`type(uint256).max` so it is not a problem to start at a non
zero nonce.