Description:
i have istio service mesh with envoy running as sidecar, a destination service with below configuration in destination rule is deployed.
mainly i was trying to check ciruitbreaker beaviour of envoy/istio.
trafficPolicy:
connectionPool:
http:
http1MaxPendingRequests: 1
tcp:
maxConnections: 5
my client is sending request to this i see a same connection limit getting applied to client side also. if a scale destination replicas to 4, would be my destination service work with total 20 connections (5 connections each pod ) ? or 5 strict connections would be shared between all 4 pods ? And how client side connection would be affected based on that ?
context:- i see after increasing the destination replicas i get less 503 with reponse flag UO.