Job configuration implementation

[gctucker]

Once we've established the principles for reworking the pipeline job configuration, this is about doing the initial implementation. It should be possible to use it locally with the kci command line tool as well as with the automated pipeline.

For example, it could be running KUnit in k8s and also in parallel run 3 kernel builds (e.g. x86_64, arm64, arm defconfigs), each of them would then be running baseline and baseline-nfs on one target platform (e.g. x86 QEMU, some arm64 Chromebook and common arm boards) for a total of 15 LAVA tests.

• minimum features implemented to have trigger / checkout / kunit / kernel-build / baseline jobs running
• ability to spawn multiple jobs in parallel with a series of parameters
• schema definitions for the new YAML job configuration created to help with validation
• tested locally with kci tool
• tested on staging