Hello all! I've been working on a little design for some improvements to the KillPolicy. You can find the design doc here <https://docs.google.com/document/d/1xRaOEe2K7OIVrDTOY9UDwwJbCIwXF3wZUrXYl8Pqy24/edit?usp=sharing> .
TL;DR: the plan is to extend the KillPolicy message to allow the initiation step of termination to be configurable. The framework can specify that a user-supplied signal be sent to initiate task termination, or the framework can supply a CommandInfo which will be executed within the task's namespaces to initiate termination. Comments on the design doc would be greatly appreciated! Cheers, Greg