As I can read here the default policy for every communication happening inside the mesh is actually retrying twice in case of, for example, 503.
I want to disable this but I cannot figure out a why to do it globally for all the mesh as Virtual Services do not allow wildcards for hosts. I don’t know if this is doable via istio manifiest or with any other CRD.
The only way I figured out is adding a Virtual Service for each K8s Service we have where I can do:
apiVersion: networking.istio.io/v1beta1 kind: VirtualService metadata: name: no-retries-for-one-service spec: hosts: - one-service.default.svc.cluster.local http: - retries: attempts: 0 route: - destination: host: one-service.default.svc.cluster.local
But this will introduce a lot of VirtualServices that we currently don’t have as we don’t use any other traffic management benefits like weighting, routing to subsets, etc.
Is this doable with only one VirtualService (or other resource) globally?