Closed ogolan-ig closed 6 months ago
🚧 This issue or pull request has been closed due to not having had activity from an Istio team member since 2024-01-31. If you feel this issue or pull request deserves attention, please reopen the issue. Please see this wiki page for more information. Thank you for your contributions.
Created by the issue and PR lifecycle manager.
Is this the right place to submit this?
Bug Description
i am experiencing unexpected behaviour regarding load balncer and http2MaxRequests . I have multiple clients and a service with replica (let's say 2 instances). I want each instance to only be able to handle one http request at a time. if the request takes some time to complete (instance is busy) the traffic should go to the other non busy instances. if all instances are busy the request should be queued.
I tried to use LEAST_REQUEST with http2MaxRequests: 1 but that caused unexpected behaviour. if one request takes some time to complete the second request returns : upstream connect error or disconnect/reset before headers. reset reason: overflow. it does not matter how many idle instances there are.
I tried using maxRequestsPerConnection: 1 , http1MaxPendingRequests: 1, maxConnections:1 and it almost worked. if one pod is busy the second request might go to the busy pod and be required or go to the idle pod. i wanted all requests to go to the idle pod.
current destination rule
Version
Additional Information
No response