feat: report action timeout as failed with timeout message

[ddneilson]

What was the problem/requirement? (What/Why)

Previously, if an OpenJD Action was canceled due to a timeout being reached then we'd report the action as just canceled. The change to update to openjd-sessions 0.5.0 ( #160 ) made it so that timeout actions would report as FAILED, but didn't change the failure message to make it clear that the reason for the failure was a timeout.

What was the solution? (How)

We mutate the action status when we recieve it to override the failure message with one that indicates that the action has reached its runtime limit.

What is the impact of this change?

The customer should have an easier time of identifying which tasks/actions failed due to a timeout rather than having just been canceled for some other reason.

How was this change tested?

I updated a unit test, and also ran the agent against the service with a sleep job that has a runtime limit. Here's a snapshot of one of the session actions that results:

% aws deadline get-session-action --farm-id $FARM_ID --queue-id $QUEUE_ID --job-id $JOB_ID --session-action-id $SA_ID
{
    "sessionActionId": "sessionaction-f7efc3852c574cc0b6fb92a33ff140d6-2",
    "status": "FAILED",
    "startedAt": "2024-02-21T20:11:15.816000+00:00",
    "endedAt": "2024-02-21T20:12:00.834000+00:00",
    "progressPercent": 73.3333,
    "sessionId": "session-f7efc3852c574cc0b6fb92a33ff140d6",
    "processExitCode": -9,
    "progressMessage": "TIMEOUT - Exceeded the allotted runtime limit.",
    "definition": {
        "taskRun": {
            "taskId": "task-0cee490179da401aa4c5f0327906eb6c-0",
            "stepId": "step-0cee490179da401aa4c5f0327906eb6c",
            "parameters": {}
        }
    }
}

Was this change documented?

N/A

Is this a breaking change?

No