Unresponsive workers should be flagged on the dashboard

[crusaderky]

• Related to worker-ttl timeout should attempt a nanny restart #8537

Let's hamstring a worker:

async def kill_event_loop():
    while True:
        pass

fut = c.submit(kill_event_loop)

This makes the worker completely unresponsive. After 5 minutes (distributed.scheduler.worker-ttl) without a single heartbeat coming through, the scheduler will disconnect it forcefully.

Until that happens, this is what I'm seeing on the dashboard:

Desired behaviour

There should be an indication on the dashboard of how many seconds have passed since the last heartbeat.
There should be also be visual cues - e.g. the whole line turning red - to indicate when you get to a significant threshold since the last heartbeat - e.g. max(4x of the expected heartbeat ratio, 25% of the worker-ttl)

[jrbourbeau]

There should be an indication on the dashboard of how many seconds have passed since the last heartbeat.
There should be also be visual cues - e.g. the whole line turning red - to indicate when you get to a significant threshold since the last heartbeat

I believe this is on the "Info" page