Fault Injection

This task shows how to inject delays and test the resiliency of your application.

Before you begin

  • Setup Istio by following the instructions in the Installation guide.

  • Deploy the Bookinfo sample application.

  • Initialize the application version routing by either first doing the request routing task or by running following commands:

    istioctl create -f samples/bookinfo/routing/route-rule-all-v1.yaml
istioctl replace -f samples/bookinfo/routing/route-rule-reviews-test-v2.yaml

Fault injection

Fault injection using HTTP delay

To test our Bookinfo application microservices for resiliency, we will inject a 7s delay between the reviews:v2 and ratings microservices, for user “jason”. Since the reviews:v2 service has a 10s timeout for its calls to the ratings service, we expect the end-to-end flow to continue without any errors.

  1. Create a fault injection rule to delay traffic coming from user “jason” (our test user)

    istioctl replace -f samples/bookinfo/routing/route-rule-ratings-test-delay.yaml

    Confirm the rule is created:

    istioctl get virtualservice ratings -o yaml

    apiVersion: networking.istio.io/v1alpha3
kind: VirtualService
metadata:
  name: ratings
  ...
spec:
  hosts:
  - ratings
  http:
  - fault:
      delay:
        fixedDelay: 7s
        percent: 100
    match:
    - headers:
        cookie:
          regex: ^(.*?;)?(user=jason)(;.*)?$
    route:
    - destination:
        name: ratings
        subset: v1
  - route:
    - destination:
        name: ratings
        subset: v1

    Allow several seconds to account for rule propagation delay to all pods.

  2. Observe application behavior

    Log in as user “jason”. If the application’s front page was set to correctly handle delays, we expect it to load within approximately 7 seconds. To see the web page response times, open the Developer Tools menu in IE, Chrome or Firefox (typically, key combination Ctrl+Shift+I or Alt+Cmd+I), tab Network, and reload the productpage web page.

    You will see that the webpage loads in about 6 seconds. The reviews section will show Sorry, product reviews are currently unavailable for this book.

Understanding what happened

The reason that the entire reviews service has failed is because our Bookinfo application has a bug. The timeout between the productpage and reviews service is less (3s + 1 retry = 6s total) than the timeout between the reviews and ratings service (10s). These kinds of bugs can occur in typical enterprise applications where different teams develop different microservices independently. Istio’s fault injection rules help you identify such anomalies without impacting end users.

Notice that we are restricting the failure impact to user “jason” only. If you login as any other user, you would not experience any delays.

Fixing the bug: At this point we would normally fix the problem by either increasing the productpage timeout or decreasing the reviews to ratings service timeout, terminate and restart the fixed microservice, and then confirm that the productpage returns its response without any errors.

However, we already have this fix running in v3 of the reviews service, so we can simply fix the problem by migrating all traffic to reviews:v3 as described in the traffic shifting task.

(Left as an exercise for the reader - change the delay rule to use a 2.8 second delay and then run it against the v3 version of reviews.)

Fault injection using HTTP Abort

As another test of resiliency, we will introduce an HTTP abort to the ratings microservices for the user “jason”. We expect the page to load immediately unlike the delay example and display the “product ratings not available” message.

  1. Create a fault injection rule to send an HTTP abort for user “jason”

    istioctl replace -f samples/bookinfo/routing/route-rule-ratings-test-abort.yaml

    Confirm the rule is created

    istioctl get virtualservice ratings -o yaml

    apiVersion: networking.istio.io/v1alpha3
kind: VirtualService
metadata:
  name: ratings
  ...
spec:
  hosts:
  - ratings
  http:
  - fault:
      abort:
        httpStatus: 500
        percent: 100
    match:
    - headers:
        cookie:
          regex: ^(.*?;)?(user=jason)(;.*)?$
  - route:
    - destination:
        name: ratings
        subset: v1

  2. Observe application behavior

    Login as user “jason”. If the rule propagated successfully to all pods, you should see the page load immediately with the “product ratings not available” message. Logout from user “jason” and you should see the ratings v2 show up successfully on the productpage web page.

Cleanup

  • Remove the application routing rules:

    istioctl delete -f samples/bookinfo/routing/route-rule-all-v1.yaml

  • If you are not planning to explore any follow-on tasks, refer to the Bookinfo cleanup instructions to shutdown the application.

What’s next