Diki
Diki a "compliance checker" or sorts, a detective control framework with pluggable rule sets. It's part of the Gardener family, but can be used also on other Kubernetes distros or even on non-Kubernetes environments, e.g. to check compliance of your hyperscaler accounts.
Diki is the Greek word for "trial". You can also memorise it as "Detective Investigation of Key Imperatives" or as GNU-style recursive acronym "Diki Investigates Key Imperatives". It's colloquially known as "Don't I Know It", which is a nice fit as well for what it does.
Important Note: This repository is in alpha stage. The API can change without any backwards compatibility.
Getting Started
Installation
TODO
Run
Most of Diki's run
configurations are provided through its config file. Options depend on the different providers and rulesets. Here are a couple of commands to get you started:
- Run all known rulesets for all known providers
diki run --config=config.yaml --all
- Run a specific ruleset for a known provider
diki run --config=config.yaml --provider=gardener --ruleset-id=disa-kubernetes-stig --ruleset-version=v1r8
- Run a specific rule defined in a ruleset for a known provider
diki run --config=config.yaml --provider=gardener --ruleset-id=disa-kubernetes-stig --ruleset-version=v1r8 --rule-id=242414
Report
Diki can generate a human readable report from the output files of a diki run
execution. Merged reports can be produced by setting the distinct-by
flag. The value of this flag is a list of key=value
pairs where the keys are the IDs of the providers we want to include in the merged report and the values are the unique metadata fields to be used as distinction values between different provider runs.
diki report output.json > report.hmtl
- Generate merged html report
diki report --distinct-by=gardener=id output1.json output2.json > report.hmtl
Unit Tests
You can manually run the tests via make test
.
Contributing
Contributions are very welcome. To learn more, see the contributor guide.
Credits
A special thank you to @vlerenc, @AleksandarSavchev and @dimityrmirchev who made significant contributions to the project before it was made available to the public.