The main reason why we decided to write the jenkins-operator is the fact that we faced a lot of problems with standard Jenkins deployment. Also, we heavily use Jenkins on large production Kubernetes clusters in a multi-tenant environment.

Some of the problems we want to solve:

volumes handling (AWS EBS volume attach/detach issue)
installing plugins with incompatible versions or security vulnerabilities
declarative configuration as code
lack of end to end tests
make Jenkins more suitable for dynamic environment and handle graceful shutdown properly
security and hardening out of the box
orphaned jobs with no jnlp connection
make errors more visible for end users

Documentation

Contribution

Feel free to file issues or pull requests.

TODO

Common:

simple API for generating Kubernetes events using one common format
~~VirtusLab docker registry~~ https://hub.docker.com/r/virtuslab/jenkins-operator
~~decorate Jenkins API client and add more functions for handling jobs and builds e.g. Ensure, CreateOrUpdate~~
documentation
~~VirtusLab flavored Jenkins theme~~
create Jenkins Jobs View for all jobs managed by the operator
~~jenkins job for executing groovy scripts~~

Base configuration:

~~install configuration as a code Jenkins plugin~~
handle Jenkins restart when base configuration has changed
~~install~~ and configure Kubernetes plugin (in-progress)
e2e pipelines using Kubernetes plugin
Jenkins hardening, disable insecure options

User configuration:

~~user reconciliation loop with CR validation~~
~~configure seed jobs and deploy keys~~
~~e2e tests for seed jobs~~
configure Jenkins authorization (via configuration as a code plugin or groovy scripts)
backup and restore for Jenkins jobs running as standalone job (AWS, GCP, Azure)
trigger backup job before pod deletion using preStop k8s hooks
verify Jenkins configuration events