Spike: Validate and persist CI job inputs

Before we continue the job inputs implementation, we need to answer some questions.

The initial implementation of job inputs is a 5 step process:

1. Allow job inputs to be defined in the CI YAML config -> ❓ Can we reuse existing logic for parsing and validation
2. Persist job inputs when a pipeline is created -> ❓ Where and how do we do this?
3. Create the DB structure for job inputs -> ✅ _Specs will be saved in ci_job_definitions and values will be saved in _[ci_job_inputs](Add database table for CI job inputs (#547439 (closed) - closed))
4. Job inputs are passed to Runner when a job is run -> ✅ Added in Add job inputs to job payload for runner (#543972 - closed)
5. Job inputs are interpolated by Runner -> 🚧 We have a PoC we can use for testing Draft: POC: Expand job inputs into the user script (gitlab-runner!5600 - closed)

This spike issue will answer the questions for steps 1 and 2 before we conduct the implementation in Persist job inputs spec in Ci::Build options field (#557879) and Add keyword to CI config for CI job inputs (#547440)