Ask HN: Resource Allocation Library for Tests
pabot (and others) allows parallel execution of tests but not allocation of resources or any intelligence.
From my experience it was always built in house with different complexities and levels of success.
What I am looking for is something that can answer the following-
1. Task/Test requests comes in asynchronously 2. Each request have a list of needed resources, for example a CPU, OS, Platform, Version of something or data availability 3. A manager has a list of agents (an agent can be physical, virtual or local) with their available resources and status 4. The manager allocate an agent and mark it as in-use 5. The manager waits for completion, collects the execution results (probably asynchronously) and marks the agent as not in use
It's not overly complicated to build one yourself but as always it's best to use something generic and tested. CI systems can be used to solve the problem but it's over engineering and complicates execution from a single point.
[0] https://news.ycombinator.com/item?id=33343420 [1] https://pabot.org/