Catching Exceptions in Individual Tasks and Restarting Them

Question

If I create a series of asyncio tasks in a top level class, all of which should essentially run forever, like so:

asyncio.create_task(...)
asyncio.create_task(...)
asyncio.create_task(...)
...

self.event_loop.run_forever()

# Once we fall out of the event loop, collect all remaining tasks,
# cancel them, and terminate the asyncio event loop
tasks = asyncio.Task.all_tasks()
group = asyncio.gather(*tasks, return_exceptions=True)
group.cancel()
self.event_loop.run_until_complete(group)
self.event_loop.close()

The above code doesn't handle the following situation, which I'm finding I need more and more, and I haven't seen an example in Googling or in the asyncio docs:

If one of the tasks fails with an exception, the exception doesn't get handled - all of the other tasks proceed, but that one task simply halts silently (other than the exception output).

So, how can I:

• Set up for the exception to be handled, so that the failure isn't silent any more
• Most importantly, restart the failed task, effectively running asyncio.create_task(...) again, just for that task? This would seem to require finding the task that received the exception in the event loop, removing it, and adding a new one - how to do that is not clear to me.
• Allow the tasks that didn't have issues to continue uninterrupted. Want to avoid any side effects of handling the task that received the exception.