asyncio.ensure_future() is not always awaited or saved

[manics]

Bug description

There are some calls to asyncio.ensure_future(...) that aren't awaited:

kubespawner/kubespawner/spawner.py

Lines 2803 to 2805 in 621be0e

	if previous_reflector:
	# we replaced the reflector, stop the old one
	asyncio.ensure_future(previous_reflector.stop())

kubespawner/kubespawner/spawner.py

Lines 3252 to 3257 in 621be0e

	self.log.error(
	"Pod %s never showed up in reflector, restarting pod reflector",
	ref_key,
	)
	self.log.error("Pods: %s", sorted(self.pod_reflector.pods.keys()))
	asyncio.ensure_future(self._start_watching_pods(replace=True))

kubespawner/kubespawner/spawner.py

Lines 3379 to 3382 in 621be0e

	self.log.error(
	"Pod %s did not disappear, restarting pod reflector", ref_key
	)
	asyncio.ensure_future(self._start_watching_pods(replace=True))

The docs for asyncio.ensure_future say the resulting Task will be scheduled, but also says

Save a reference to the result of this function, to avoid a task disappearing mid-execution.

asyncio.create_task() has a longer explanation

Important. Save a reference to the result of this function, to avoid a task disappearing mid-execution. The event loop only keeps weak references to tasks. A task that isn’t referenced elsewhere may get garbage collected at any time, even before it’s done.

I think we should either await those, save a reference to them, or add a comment to the code explaining why it isn't necessary.

[manics]

There's also some calls in proxy.py

kubespawner/kubespawner/proxy.py

Lines 358 to 360 in 7ba57a8

	asyncio.ensure_future(self.ingress_reflector.start())
	asyncio.ensure_future(self.service_reflector.start())
	asyncio.ensure_future(self.endpoint_reflector.start())

[consideRatio]

@manics thanks for investigating this thoroughly -- I agree on the decision to await things now that I've seen the motivation in this issue!