roelvanduijnhoven
commented
6 years ago Thanks for the PR @MatthiasKuehneEllerhold. Code looks great. I have two points that really both revolve around the name of the strategy.
- I initially thought that this PR would kill a job after some amount of time. But I now see that is instead working on the worker level.
- Also this strategy does not guarantee that the worker dies within the specified amount of time; as a job that was ran just before the deadline may happy live on.
Not sure what name would better reflect this. Curious to hear what you think @MatthiasKuehneEllerhold . What about StopWorkerAfterTimeStrategy?
The MaxRunsStrategy does not have this ambiguity as time is not really
May I ask what company you are using this repo for? Just curious :).
basz
MatthiasKuehneEllerhold
This PR adds a new strategy for terminating workers after a set amount of time.
We need this functionality in order to deal with some external interfaces and their timeouts out of our control. We dont want to have 1 endless running worker but instead have workers that run for a set amount of time and respawn.
This strategy is off by default (in order to keep BC) and is totally optional. Reasonable default values (according to our interpretation) are set (1 hour timeout) and can be overwritten with the archicture thats already in place. We figured we're not the only ones needing this kind of strategy, what do you think?
Documentation and Unit-Tests are in this as well.