Test to verify if filtering by merkleRoot works

[mateusz-sekara]

Motivation

We've implemented Solution 2 from this #726 and it caused problems during the reorgs so it was reverted #1155

Caching block_timestamps of the CommitRoots that are not finalized is extremely tricky because during the reorg CommitReport can come back to the logpoller with a different timestamp. This causes issues when some of the CommitReports might be forever skipped because the cached timestamp is higher than a new value. Keeping consistency between the cache and LogPoller would be an overkill so we need to pursue an alternative solution.

Solution

After adding DSL to the LogPoller, Solution 1 #726 sounds like a very reasonable way to go. It doesn't require any changes to LogPoller, it could be implemented using the FilteredLogs and it's completely immune to all reorgs - we move in-memory skipping snoozedRoots to the database level. Therefore, we keep fetching a reasonable number of logs to the memory. It works only on finalized logs so no need to sync any state between LP and cache. Also, the snoozed roots cache is very simple again, I have to admit that #726 led to a very complex solution at some point.
On the other hand, it's critical to load test that change to verify how filtering out snoozed roots on the db level behaves.

Based on the &726 we want to pass executed roots to the database query to fetch only unexpired roots from the database. Unexpired roots can be determined from the snoozedRoots cache. This is probably the most accurate approach because it always fetches the minimal required set from the database.
You can think of the query like this

select * from evm.logs
where address = :addr
and event_sig = :event_sig
and substring(data, X, Y) not in (:executed_roots_array)
and block_timestamp > :permissionlessThresholdWindow

[0xnogo]

nit/ we could also verify we are getting the expect log by checking on the merkleRoot