New Features & Enhancements

• Introduces azure-disk-loss experiment, which detaches the virtual disk from an Azure instance

• Introduces pod-network-partition experiment, which blocks Ingress and Egress traffic of the target application. It provides the option to specify the pod-selector, namespace-selector, and ports.

• Introduces the ability to provide default application status checks as a tunable for the pod-specific experiments from the ChaosEngine. It is helpful for the scenarios where we don’t want to validate the application status as a mandatory check during pre & post chaos.

• Enhance the GCP VM Instance Stop experiment for VMs that are a part of an Autoscaling Group, where the VMs are now awaited to fully stop before proceeding with the experiment.

• Enhancements for the cmdProbe to run the probe pod with the host network.

• Enhance probe pod functionality where the experiment serviceAccount is now passed to the probe pod. It is helpful for the scenarios where the probe pod needs specific RBAC permissions for the execution.

• litmuschaos_awaited_experiment metric contains an injection_time label. Adds injection_time label as a tunable via the INJECTION_TIME_FILTER ENV. Which can be used for the annotations inside the interleaving dashboards

• Enhance the chaos-scheduler to specify the minutes and hours separately. This provides the ability to schedule the chaos at the nth minutes per x hours. Which will be helpful for the staggered schedules to avoid collisions.

• Adds the ability to search experiments based on certain keywords in Chaos Hub. Added all the related keywords for individual experiments.

• Improves the error handling in the chaos-operator for chaosengine status nil check

• Removes the init container which modifies the permissions of the docker socket path, and instead provides the docker socket path inside the DOCKER_HOST ENV for the Pumba Lib.

• litmus metrics are now exported only for the active ChaosEngines. It stops overriding the value of existing metrics as soon as ChaosEngine comes to the completed state.

• Enhance the e2e pipeline by adding debug steps and using shared runners to run tests in parallel, also allow the ability to run a test with a custom service account to validate a specific set of permissions required for an experiment.

Major Bug Fixes

• Fixes chaos-result name generation where the experiment-name and instance-id are passed inside the helper pods, which are used to generate the chaos-result name.

Installation

kubectl apply -f https://litmuschaos.github.io/litmus/litmus-operator-v2.1.0.yaml

Verify your installation

• Verify if the chaos operator is running
  kubectl get pods -n litmus

• Verify if chaos CRDs are installed
  kubectl get crds | grep chaos

For more details refer to the documentation at Docs