Auto-recover IP addresses by running autodiscovery at a certain interval

[cerahmed]

Is your feature request related to a problem? Please describe.
I have more than 10 AC units to manage. I use HA's default Gree Climate integration, and I'm experimenting with this custom integration to find the pros and cons of each.

This integration is better all-around, for its ability to expose vertical/horizontal swings, and has a beeper control. However, one feature it's missing that HA's built-in integration has is it does not require static IP's to work.

HA's integration checks for IP Addresses changes every 10 minutes, by rediscovering network devices, comparing to already registered devices, then updating the IP addresses as needed.

Describe the solution you'd like
Allow the autodiscovery to re-run after a set amount of time, and compare the mac addresses/ip addresses to already registered devices.

Describe alternatives you've considered
Setting static IP's for each individual AC unit is still possible but time consuming and unnecessary for most users.

[p-monteiro]

I think network conditions shouldn't change every 10 minutes, maybe that is too much of a frequency?
Would this being performed during integration initialization be acceptable for you?

[cerahmed]

Hey @p-monteiro, thanks for your response.

I agree with you that a 10-minute interval isn't necessary, knowing that most networks default to having a lease time of 1 day. However the good part is that it limits device downtimes to a maximum of 10 minutes - a balance between not being user noticeable and processing effeciency.

I just looked at HA's gree component, they actually perform the discovery scan and update entities' IP addresses as needed every 300 seconds (5 minutes), which in practice isn't processor taxing and I haven't noticed any downtimes during months of use.

Being performed during integration initialization (after a reboot for example) is definetely better than having to reconfigure AC units (or having to setup static ip's), but on a longer system uptime it would still cause some inconveniences.

I'd be glad to help implement/test this feature request with my moderate skills.

[p-monteiro]

@RobHofmann I implemented this in my PR (here)

The check is made when a connection fails in the coordinator and when the integration starts, so no periodic scan interval.

[cerahmed]

Hey @p-monteiro .. thanks for considering this enhancement!