If we are going to speculate, perhaps the reason this is taking so long is because the release is larger than usual, and/or this is vacation time...
Or maybe it's taking a while because the next release really matters.
I should not have said that...
Or, (throwing @JasonJoel a bone) maybe they are finally following up on long range plans.
Just can't help myself...
I heard (because I just thought it in my head) that the next version is going to be 3.0.1. YMMV.
You are even factoring a fix into your dream version.... How pragmatic of you...
It's really, really quiet, they must be working on something big...
Yeah, not even a beta for 2+ months. Here's my wishlist for 3.x or 2.4.x:
- Memory Fix (Yes, this has become critical for me. Rebooting everything 3 days sucks)
- Better memory leak tools to detect where the problem is (I have a theme here)
- HomePod TTS. Then I can jettison Echo Speaks which my gut says is a huge resource drain but too hard to live without
And, yes, I know that splitting the load between two hubs would likely help. But, every time I look at the migration process I am overwhelmed with all the problems I can see happening. because I use Amazon Echo Skill, Bond, Echo Speaks, Ecobee, HomeBridge, IFTTT, Kasa and LifX. Yes, it's a lot of LAN and Cloud based add ons. But they worked just dandy for years until recently.
Basically, I'm hoping the next release is a silver bullet. Because lately I've been pining away for SmartThings and I NEVER thought I'd say that. OK that went too far. Strike the last line about ST.
Help me HE developers, you're my only hope.
You need to find your cause first. It's unlikely anything will be fixed if they can't target it
Bryan talked about it. Hopefully by the end of the year.
Better memory leak tools to detect where the problem is
Hence the reason for bullet two.
You cannot really tell what is taking memory and not giving it back from the two things I can think of to look at:
I guess I'm going to completely eliminate Echo Speaks. I've been working on refactoring everything that uses it to be a "routine" in Alexa. It's not that easy though as things that are really clever like that I can code in RM are just not possible in Alexa Routines. And you can't use variables and device values in phrases. The things that made HE seem magical just doesn't work anymore. Yes, I'm whining. But, they really don't give you the tools to know WHY the free memory just goes down, down, down until poof. It's not backups like others claim. I can't find a correlation there.
And I have been trying. Been reducing logging. Eliminated rules altogether. Rewriting rules to make them simpler. Removing devices that are not currently in use. It's all shooting in the dark.
It's not very hard really. Take a Local Backup of your hub. Then go through all the Apps you mentioned and disable them. On the second hub, restore the backup and disable everything that is still enabled on the first.
That will get probably 80%. Some clicky is needed to "renew" connections in some cases. But as a quick test to see if there are advantages, it's hard to get simpler/easier. You could probably do the whole proof-of-concept in an hour. The next step is to use HubMesh to mirror the devices back to the original hub. Here's where Swap App Devices is golden.
If you wish to delve deeper into that, start a new Topic
I think it was more a case of the cart before the horse in the ordering of @wiegout 's list, you are right you need to find the cause, but for that some more visibility of what's happening would help.
And I get you are offering an option to troubleshoot the situation @csteele , which is great, but I think it would not be something everyone could do, i.e. not everyone has a second hub. More options to dig into what's happening would always be welcomed through built-in features.
Exactly. Thanks. Need visibility thatβs not currently available.
Been trying to find a time when my wife won't notice that things aren't working. When things don't behave as normal she loses confidence in the entire platform. I know you all know what I mean. I try to make it reliable as dial tone (back in the day.)
I get that but sometimes diagnostics can be a pain. I mean look at it this way. How do you know it's not the alexa integration unless you can turn it off to observe the effect either positively or negatively? Same goes for anything else. At some point you need to bite the bullet to forge ahead. Even in medicine, sometimes the cure is extremely painful but worth it in the end....
"Rip the Band-Aid off"
I was able to find what was causing my freeMemory loss, but first, I explained to my wife that the only way for me to figure it out was to take one thing away every couple of days. Losing a little functionality in steps is better than having a non-responsive hub after four days. It also provided an opportunity to re-evaluate what we want and don't want out of our smart home.
I just had 29 days of uptime and for the last week freeMemory had been staying at 360-375MB. I only rebooted because I added the offender back to the hub for final verification.
We're all dying to know what it was.
For me "Splitting the Load" was a fallacy. Hub Mesh was the cause.
FWIW i moved my LAN based items (Kasa, Switchbot, ESPhome, Weather) to my C7 and it really reduced the memory load on my C8.
I've also been noticing the lack of updates and staff forum activity lately. Hopefully everything is ok