Currently, Gridsync depends on running tahoe and magic-folder subprocesses for its core folder-syncing functionality and will spin up, parse the outputs of, and shut down these processes as needed throughout the course of normal operation. If one of these subprocesses is terminated by a cause that is "external" to Gridsync, however (for example, due to an OOM issue, because the user killed it, etc.) Gridsync will continue to run as though the processes were still available -- waiting for outputs from various "status" endpoints that, in absence of the corresponding running tahoe/magic-folder subprocess, will never return, leaving the application to essentially idle forever.

Instead, Gridsync should take additional measures to detect when one of its tahoe/magic-folder subprocess has become unavailable and take appropriate steps as needed given the circumstance (for example, by relaunching processes that may have crashed) to ensure that the application continues to operate normally.