adding retryon support

Signed-off-by: brandoncate <brandon.cate@shipt.com>

pkg/apis/projectcontour/v1/httpproxy.go

@@ -255,6 +255,10 @@ type TimeoutPolicy struct {
 	Idle string `json:"idle,omitempty"`
 }
 
+// RetryOn is a string type alias with validation to ensure that the value is valid.
+// +kubebuilder:validation:Enum="5xx";gateway-error;reset;connect-failure;retriable-4xx;refused-stream;retriable-status-codes;retriable-headers;cancelled;deadline-exceeded;internal;resource-exhausted;unavailable
+type RetryOn string
+
 // RetryPolicy defines the attributes associated with retrying policy.
 type RetryPolicy struct {
 	// NumRetries is maximum allowed number of retries.
@@ -264,6 +268,28 @@ type RetryPolicy struct {
 	// PerTryTimeout specifies the timeout per retry attempt.
 	// Ignored if NumRetries is not supplied.
 	PerTryTimeout string `json:"perTryTimeout,omitempty"`
+	// RetryOn specifies the conditions on which to retry a request.
+	//
+	// Supported [HTTP conditions](https://www.envoyproxy.io/docs/envoy/latest/configuration/http/http_filters/router_filter#x-envoy-retry-on):
+	//
+	// - `5xx`
+	// - `gateway-error`
+	// - `reset`
+	// - `connect-failure`
+	// - `retriable-4xx`
+	// - `refused-stream`
+	// - `retriable-status-codes`
+	// - `retriable-headers`
+	//
+	// Supported [gRPC conditions](https://www.envoyproxy.io/docs/envoy/latest/configuration/http/http_filters/router_filter#x-envoy-retry-grpc-on):
+	//
+	// - `cancelled`
+	// - `deadline-exceeded`
+	// - `internal`
+	// - `resource-exhausted`
+	// - `unavailable`
+	// +optional
+	RetryOn []RetryOn `json:"retryOn,omitempty"`
 }
 
 // ReplacePrefix describes a path prefix replacement.

pkg/apis/projectcontour/v1/zz_generated.deepcopy.go

@@ -300,6 +300,11 @@ func (in *ReplacePrefix) DeepCopy() *ReplacePrefix {
 // DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.
 func (in *RetryPolicy) DeepCopyInto(out *RetryPolicy) {
 	*out = *in
+	if in.RetryOn != nil {
+		in, out := &in.RetryOn, &out.RetryOn
+		*out = make([]RetryOn, len(*in))
+		copy(*out, *in)
+	}
 	return
 }
 
@@ -338,7 +343,7 @@ func (in *Route) DeepCopyInto(out *Route) {
 	if in.RetryPolicy != nil {
 		in, out := &in.RetryPolicy, &out.RetryPolicy
 		*out = new(RetryPolicy)
-		**out = **in
+		(*in).DeepCopyInto(*out)
 	}
 	if in.HealthCheckPolicy != nil {
 		in, out := &in.HealthCheckPolicy, &out.HealthCheckPolicy

pkg/router/contour.go

@@ -19,6 +19,7 @@ package router
 import (
 	"context"
 	"fmt"
+	"strings"
 
 	"github.com/google/go-cmp/cmp"
 	"github.com/google/go-cmp/cmp/cmpopts"
@@ -444,11 +445,22 @@ func (cr *ContourRouter) makeRetryPolicy(canary *flaggerv1.Canary) *contourv1.Re
 		return &contourv1.RetryPolicy{
 			NumRetries:    uint32(canary.Spec.Service.Retries.Attempts),
 			PerTryTimeout: canary.Spec.Service.Retries.PerTryTimeout,
+			RetryOn:       makeRetryOn(canary.Spec.Service.Retries.RetryOn),
 		}
 	}
 	return nil
 }
 
+func makeRetryOn(retryOnString string) []contourv1.RetryOn {
+	retryOnSplit := strings.Split(retryOnString, ",")
+
+	retryOn := make([]contourv1.RetryOn, len(retryOnSplit))
+	for i, v := range retryOnSplit {
+		retryOn[i] = contourv1.RetryOn(v)
+	}
+	return retryOn
+}
+
 func (cr *ContourRouter) makeLinkerdHeaderValue(canary *flaggerv1.Canary, serviceName string) contourv1.HeaderValue {
 	return contourv1.HeaderValue{
 		Name:  "l5d-dst-override",

pkg/router/router_test.go

@@ -132,6 +132,7 @@ func newTestCanary() *flaggerv1.Canary {
 				Retries: &istiov1alpha3.HTTPRetry{
 					Attempts:      10,
 					PerTryTimeout: "30s",
+					RetryOn:       "connect-failure,gateway-error",
 				},
 				Gateways: []string{
 					"public-gateway.istio",

test/contour/test-custom.sh

@@ -0,0 +1,57 @@
+#!/usr/bin/env bash
+
+cat <<EOF | kubectl apply -f -
+apiVersion: flagger.app/v1beta1
+kind: Canary
+metadata:
+  name: podinfo
+  namespace: test
+spec:
+  provider: contour
+  targetRef:
+    apiVersion: apps/v1
+    kind: Deployment
+    name: podinfo
+  progressDeadlineSeconds: 60
+  service:
+    port: 80
+    targetPort: 9898
+    retries:
+      attempts: 3
+      perTryTimeout: 5s
+      retryOn: connect-failure,5xx
+  analysis:
+    interval: 15s
+    threshold: 15
+    maxWeight: 50
+    stepWeight: 10
+    metrics:
+    - name: request-success-rate
+      threshold: 99
+      interval: 1m
+    - name: request-duration
+      threshold: 500
+      interval: 1m
+    webhooks:
+      - name: acceptance-test
+        type: pre-rollout
+        url: http://flagger-loadtester.test/
+        timeout: 10s
+        metadata:
+          type: bash
+          cmd: "curl -sd 'test' http://podinfo-canary/token | grep token"
+      - name: contour-acceptance-test
+        type: pre-rollout
+        url: http://flagger-loadtester.test/
+        timeout: 10s
+        metadata:
+          type: bash
+          cmd: "curl -sd 'test' -H 'Host: app.example.com' http://envoy.projectcontour/token | grep token"
+      - name: load-test
+        url: http://flagger-loadtester.test/
+        timeout: 5s
+        metadata:
+          type: cmd
+          cmd: "hey -z 1m -q 5 -c 2 -host app.example.com http://envoy.projectcontour"
+          logCmdOutput: "true"
+EOF