Manager has limits on it already -- currently has guaranteed QoS.

Related to #694 and probably a pre-req -- set a small request limit to give the workspace pod burst-able QoS.

Additional considerations:

• use GOMEMLIMIT to tell the agent to not use more than the requested memory for itself. Consider setting it using the Downward API. Avoid setting it on the child processes (use SetMemoryLimit in code?).
• check for zombie sub-processes.