This PR improve the handling of communication failure
It adds the ability to poll a device - although this isn't currently used
Devices that have failed are reported as should_poll since https://github.com/pavoni/pyvera/issues/147 highlights that the subscription mechanism doesn't handle recovery reliably.
I investigated forcing a poll when refreshing a failed device - but it seemed to cause a lot of zwave traffic - and stop the normal services working! I think it need to be done far less often, with more care. This requires more careful thought.
This PR improve the handling of communication failure
should_poll
since https://github.com/pavoni/pyvera/issues/147 highlights that the subscription mechanism doesn't handle recovery reliably.I investigated forcing a poll when refreshing a failed device - but it seemed to cause a lot of zwave traffic - and stop the normal services working! I think it need to be done far less often, with more care. This requires more careful thought.