Chaos Dashboard is the user-facing dashboard for Harness Chaos Engineering, which provides access to its different features. These features can be broadly classified into:
- Experiment Management
- User Management
- Chaos Infrastructure Management
- Chaos Analytics
Experiment management is an umbrella term for all the actions related to a chaos experiment. This includes their creation, update, and deletion among other things.
- Creation of chaos experiments can be done from:
- Blank Canvas
- Experiment Templates (From Chaos Hubs)
- Experiment Manifest File
- An existing chaos experiment can be updated to make changes to the existing experiment or to a copy of the experiment.
- Deleting a chaos experiment removes it from the dashboard.
Role Based Access Control can be enforced for the users to assign them a role, such that only the requisite resources defined under that role can be accessed by them.
- Custom roles can be created by selectively providing access to any set of resources from the list of all available resources.
Users can also be grouped so as to provide them a common role or adding them to a common notification channel, etc.
Chaos Infrastructure Management
Chaos infrastructures can be added to or removed from the environments.
- Kubernetes chaos infrastructures can be installed in either cluster-wide scope or namespace mode, to enable chaos injection through all namespaces or only a single namespace respectively.
- All the chaos infrastructure services adhere to the principle of least privilege.
- Multiple chaos infrastructures may be added to a single environment.
- All the chaos infrastructures added under a given environment can be targeted using a chaos experiment.
As part of a chaos experiment run various statistics are obtained, such as:
- Resiliency Score: A quantitative measure of the resiliency of the target environment for the given experiment.
- Average Resiliency Score: Average of resiliency scores obtained over consecutive experiment scores.
- Probe Success Percentage (Per Fault): The percentage of successful probe checks out of the total number of probes defined for a fault.