Have anyone gotten this error, like timeout around 30 sec for a long running graphql query? I was using graphiql to retrieve lineage results of about 3000 I suspect is the graphql servlet, but am unsure how to set the timeout to be longer?

{
  "servlet": "apiServlet",
  "message": "Service Unavailable",
  "url": "/api/graphql",
  "status": "503"
}

hey Aezo! hmm if you're only expecting about 3,000 results I would expect that 30 seconds should be enough time to return what you would expect. Are any of your pods/containers unhealthy? especially GMS and elasticsearch

Hmm, our elasticsearch is managed service. I think its heavy in the sense that its searching through the lineage. Gms replicas is currently 3-5 . Is the 30 seconds configurable?

Currently it's not... But either way taking 30s on a query is not ideal at all. It certainly should not take this long for 30 lineage

I think it have to search through all the entities that have lineage link to dataset… so although it might return just 3000 records. It maybe be check 130k ++ datasets. I dont think bumping max bucket can always be the solution too. 🥲. Not sure if ES or servlet got timeout is the true cause.