We are currently implementing Test Execution of Test Suites using Gitlab's CI.
One of the issue we encountered is that it keeps using a runner even though it is being used by another job.
Problem 1:
- There are 3 available runners (different machines), config is already set to limit = 1 and request_concurrency = 1.
- Pipeline with 3 jobs running in parallel will be triggered.
Result: The 2 jobs will use 1 runner.
Expectation: The 3 jobs will use the 3 runners (1 job per runner).
Problem 2:
- Given the above conditions.
- Pipeline with 3 jobs running in parallel will be triggered twice.
Result: almost all jobs are running, executing multiple jobs in a runner.
Expectation: First pipeline should be running (1 job per runner) and second pipeline should be pending.
Second pipeline's job/s will only be executed once done with first pipeline's job/s