Join GitHub today
GitHub is home to over 31 million developers working together to host and review code, manage projects, and build software together.Sign up
balance/recover the load distribution when new slave joins #970
With Locust master and slave agents running in Kubernetes, Kubernetes will guarantee the availability of Locust agents.
But when a slave agent crashes and restarts, it will have a different client id and it has no idea of the user load that master assigned to it previously. And the total number of running locusts will not be as many as expected.
So it might be a better way to balance the user load when new client joins, and the total number of running locusts will still be the same as we specified in the swarm request.
also this PR fixes some issue I noticed when running in Python 3 with web mode, it turns out to be the inconsistency introduced in recv_from_client and send_to_client
Any thought or comment?