Connect version 4.20 broke local http access?

I agree with Jan - with this and other Connect regressions, as well as other CIQ networking limitations that prevent me from providing real functionality, its clear Garmin isn't serious about CIQ and I have adjusted the effort I put it to apps accordingly.

Thanks a lot for the provided info Brandon, although this "not knowing" is far from happy news. 

With this approach seen also on past resolutions of other issues related to the CIQ platform, Garmin sends clear message how insignificant is Connect IQ for the company and how we developers will be treated even for the future. Obviously CIQ didn't grow fast enough for your managers - we can only debate what was the primary cause. I decided to invest lots of my energy (and money) recently to try to bring more of stronger outdoors companies to your platform, but not only number of issues along the development of the app for the first one is surprising, but this current issue - genially "planned" to cover most of of summer prime-time - is probably not something from what I will recover to continue with other negotiated companies. One would say this is not exactly on par with your strategy... 

So I keep asking what are the actual goals for CIQ platform and how Garmin organisation structure look like. Because from outside it seems that manager of one team is prohibited to talk to manager of second team and do some effective decisions to move things further. I was part of such company once, and it was sad experience. 

This is just simple summarization how I feel as developer of your platform when I am not trying to implement simple watchface or another basic-formula datafield. And I am not the only one.

BTW, if anybody manage to have local connectivity via https from Android, let me please know - I already gave up. 

Thanks for your engagement!

I don't have any specific updates on this yet, but it's being looked at by the GCM team. I think there are a lot of strong use cases mentioned in this thread, and the suggestion that non-routable addresses should be available over HTTP seems entirely reasonable to me.

The earliest that this could be addressed is in the next GCM release, which should be later this month (GCM is on a monthly release cycle), but I can't say whether that's a realistic expectation. We're doing what we can to keep the GCM team aware of the impact to all of you and get decisions made. I apologize to everyone for the road block this has imposed. While jim_m_58 is correct that we advised switching to HTTPS in 2017, this clearly hasn't been strictly enforced and we apparently had a blind spot with regard to some of these use cases that require HTTP.

I'm optimistic that we'll have a solution, but it's unclear to me the time frame required to get one in place.

I can not overstate how much of a god-send this capability is. I have become very reliant on this capability. If this localhost issue isn't resolved I will be staying on 4.19 for as long possible.