Requirements gathering for Application Resiliency assessments
michaelshire opened this issue · comments
Current goal is to evaluate all workloads in a project/namespace and determine a result and map to a colour:
- Workload Name
- Workload Type [Deployment, DeploymentConfig, StatefulSet, DaemonSet, CronJob]
- Strategy is "RollingUpdate" (false=red, true=green)
- Resources Limits cpu are defined (false=red, true=green)
- Resources Limits memory are defined (false=red, true=green)
- Resource Requests cpu are defined (false=red, true=green)
- Resource Requests memory are defined (false=red, true=green)
- Liveness checks are defined (false=red, true=green)
- Readiness checks are defined (false=red, true=green)
- A poddisruptionbudget is defined that matches the workload (false=yellow, true=green)
- A horizontal pod autoscaler has been defined that matches the workload (false=yellow, true=green)
If any boolean cases the result is "false", provide an explanation and actions to remediate.
Currently working in python to capture and present the findings.
Script can easily be run against all projects with requester=application-registry and emailed results to the project's metadata.annotations.contacts.
Looking for input on next steps.
@michaelshire can you please estimate this issue?
Presented to Justin & Olena October 20