Fix storage node syncing query#4621
Conversation
|
The latest updates on your projects. Learn more about Vercel for Git ↗︎ 1 Ignored Deployment
|
mnaamani
left a comment
There was a problem hiding this comment.
Fix looks correct, serially making multiple queries is fine. Would doing multiple queries in parallel give a faster response?
zeeshanakram3
left a comment
There was a problem hiding this comment.
Thanks for the fast fix!
- I think we should add the
onErrorerror handling Apollo link while initializing the Query-node apollo client in storage node, similar to how it's implemented in the distributor-node's QueryNodeApi. This will imporve error handling/logging for graphql syntax & network errors - Should not be in the scope of this PR I think, but we should analyze all the queries having structure similar to storageDataObjectsConnection that are being used in all of the services/nodes, i.e. queries having one or more array params. And then if there is a chance that those queries could too break the default request size limit, we should split those queries too into smaller ones
Added in fa7cfb3
Created a new issue: #4624 |
Addresses #4615 by splitting
getAllAssignedDataObjectsquery into smaller queries, so that no individual query has more than 1000 storage bag ids provided in the input (to avoid hitting the 100kb request size limit)