Improve the way we track job identities.

[ispasov]

Description

• Improve the way we track job identities.

JobID was used to uniquely identify jobs.
This ID used to allow us to cnacel a runner provisioning if the job is cancelled.
However, there are cases in which the JobID is not avaialble, but the RunnerRequestId is.

There are two main flows:

1. A job is assinged to the integration
   • In this case there is a JobID, but the RunnerRequestId is not available.
2. A job is acquired by the integration
   • In this case there is a RunnerRequestId, but the JobID is not available.

We now use a jobIdentity struct to uniquely identify a job.
This struct contains the JobID and the RunnerRequestId.

The decision to cancel a runner provisioning is now based on the jobIdentity struct.
In the case both IDs are not available, we will not cancel the runner provisioning.

Testing

Standard Run

Run a job and ensures it completes

Cancel Job

Ensure there are not enough resources on an env
Start a job
See that provisioning starts to fail
Cancel the job
Ensure the provisioning stops

Multiple Jobs

Ensure there are not enough resources on an env
Start a job
Start another job
Cancel one of the jobs
Free space in the env
The non-cancelled job should complete