Closed relyt29 closed 1 month ago
It seems like we could have maintained the parallelization just by adding Mutex on specific keeper functions
I'm actually going to close this in favor of #458 - when 458 is merged, this issue becomes moot because the block proposer no longer has to fire off these kinds of requests to the off-chain components
What is the purpose of the change
this change removes parallelization in the topics handler over topics. Instead of parallelizing the math to craft the request for each topic, this change only allows the actual submission of HTTP requests to be parallelized.
The reason for this is that it makes the code simpler to reason about especially with regards to race conditions across many threads accessing the same keeper
Testing and Verifying
This code is covered by our existing unit tests. No functionality is changed.
Documentation and Release Note
This PR has no implications for our documentation