perf: remove blocking calls from POI queries #6300
Conversation
MoonBoi9001
force-pushed
the
feat/async-poi-resolver
branch
from
8045272 to
b6cd094
Compare
Remove synchronous `block_on` call in `resolve_proof_of_indexing` which was blocking tokio worker threads while waiting for database queries. Before this change, each POI query blocked an entire tokio worker thread. After this change, POI queries properly yield to the async runtime while waiting for database I/O, allowing the connection pool to be fully utilized. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Change the publicProofsOfIndexing resolver to process all POI requests in parallel using `future::join_all` instead of sequentially in a for loop. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
MoonBoi9001
force-pushed
the
feat/async-poi-resolver
branch
from
b6cd094 to
888a12e
Compare
|
The first commit is great and something that absolutely needs to happen. For the second commit, I am not quite clear on what the motivation is - are PoI requests slow? The worry I have is that when |
|
The motivation is POI verification tooling for dispute investigation. When investigating POI discrepancies (e.g., for arbitration disputes), we need to compute POIs across large block ranges to identify where divergence occurred. With sequential processing, throughput is bottlenecked at ~2k POIs/second even though the database connection pool has capacity. The publicProofsOfIndexing endpoint accepts batch requests (up to 10), but processing them sequentially means we're not utilizing available parallelism. Checking every block sequentially on a network with 100m blocks takes 13+ hours at ~2k blocks/second. |
Pre-fetch all block hashes in a single batch query before parallel POI processing, reducing database round-trips from 10+ to 1-2 per batch. - Add block_hashes_by_block_numbers batch method to ChainStore trait - Add get_public_proof_of_indexing_with_block_hash to StatusStore trait - Modify resolver to group requests by network and batch-fetch hashes - Pass pre-fetched hashes to avoid redundant lookups during parallel POI Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
MoonBoi9001
force-pushed
the
feat/async-poi-resolver
branch
from
8b2a3fc to
4f89b3f
Compare
The third commit addresses this, now pre-fetch all block hashes in a single batch query before parallel processing, reducing DB round-trips. |
Summary
Two optimizations for POI query handling in the index-node resolver:
Make
resolve_proof_of_indexingasync - Removes synchronousblock_oncall that was blocking tokio worker threads while waiting for database queries. POI queries now properly yield to the async runtime.Parallelize
publicProofsOfIndexingrequests - Changes sequential for-loop tofuture::join_allfor concurrent execution of batch POI requests.Changes
server/index-node/src/resolver.rsresolve_proof_of_indexing:fn→async fn,block_on(fut)→fut.awaitresolve_public_proofs_of_indexing: sequential loop →future::join_allfutureimport fromgraph::futures03