Add additional timestamp fields

We need to track more events during some of the jobs:
1)How long task spend in "queued" state (as k8s cluster might be too busy), so we need ts_start_queued, ts_start_running
2)Same for tests, DUT might not be immediately available, and task will stay in queue

So total ts required: queued, running, completed
Will help also to estimate ETA when task start running.

No files or tests are named. Start by tracing where job and test states move from queued to running to completed, including waits for the Kubernetes cluster or DUT; done means recording the requested timestamps and using them to measure queue duration and estimate ETA.

Written by the indexing model from the issue text.