[CI] Investigate why a sleep is needed after adding a runner for it to start accepting jobs

[anjannath]

While working on #282 we found that scheduling a job immediately after adding the self-hosted runner, the runner would not accept the job, this was worked around by adding a sleep to create an artificial delay between adding the runner and submitting the first job for that runner, we should investigate why this delay is needed.

mapt/.github/workflows/provision-hosted-runner.yaml

Lines 60 to 61 in 7d5d5e1

	- name: wait for runner to start accepting jobs
	run: sleep 120