Define Resilience Policies
Resilience policies are a cornerstone of Steadybit's capabilities, and in this guide, you will learn what they are and how to use them.
You can think about resilience policies like linting or security rules (ESLint, SonarQube, dependency auditing and more) for your systems' resilience. Resilience policies help you identify issues and (non-) compliance with the desired expectations. For example, Steadybit can outline that a particular Kubernetes configuration best practice isn't followed or that an HTTP request is missing timeouts or retries. Resilience policies are declarative by nature. This means that resilience policies are easy to get started with. No need for a resilience or chaos engineering expert!
You can get started with resilience policies via our CLI or the UI. This guide will help you get started with our CLI, which has greater flexibility and ensures that you can apply configuration as code best practices.
npm install -g steadybit
Once the CLI is installed, you will find that a new global executable named
steadybitis available. Before you can start using policies, you need to authenticate with your Steadybit account via an API access token. Authentication profiles are the most comfortable way to do so:
steadybit config profile add
Resilience policies are used via so-called policy bindings. When using the CLI, you will be interacting with policy binding YAML files (
.steadybit.yml). To create such a file, you can use the following command:
steadybit policy-binding init
You can use the following command to upload the policy binding file to your Steadybit platform. Don't worry; this won't execute any attacks or experiments.
steadybit policy-binding apply
Now that the Steadybit platform knows about your policy binding and its expectations via a linked policy, it is time to check how many of the selected tasks are fulfilled!
steadybit policy-binding verify
Depend on your selection in step 3, you may find that not all tasks are fulfilled! To learn more about each tasks' state you can open the Steadybit UI to learn more and to execute experiments!
steadybit policy-binding open
Image showing Steadybit's landscape view with a selected Kubernetes deployment that has fulfilled 40% of its policies
You can also execute the linked policies of your policy binding from the CLI via:
steadybit policy-binding exec
This will execute all experiments and weak spots and checks whether your system is compliant to it.
Within your CI/CD the
--yes-parameter comes in hand to skip the confirmation dialog.
This guide has shown how to use the CLI to declare a service's desired resilience. Approaching resilience in a declarative way makes it easier to get started. On top of this, we leveraged a configuration as code approach which helps us evolve, version and update our policies together with the described service. There was no need to know Kubernetes best practices or chaos engineering specifics. Resilience policies enable everyone to rely on reusable (pre-defined) policies and tasks to observe and improve service resilience!