Correct me if I’m wrong but the Wemos D1 runs on 5VDC, or at least mine have micro-USB and it can be fed over this.
Yes, it does have onboard 5 to 3.3V regulator, so you can omit that in converter design. So, for testing only program esp, connect Rx, Tx and see if it connects. Pull-ups are supposely in most of the times unneeded (again, try and see).
Is there a way to slow down queries to melcloud in HA?
It’s not just that. It happened 2 or 3 times since i have mitsubishi climates that they sort of “changed key” (or similar) so i’ve had to remove integration and re-add it to renew this key. It seems that melcloud app does this seamlessly, while HA just stop working. And, of course, this always happened when i was away so i couldn’t turn on cooling to come into a cool house.
I have the same problem: the MELCloud integration stopped working at 1700 yesterday when I updated to HA OS 11.5, along with my REST sensors. The Android app & web interface to MELCloud all report this same “excessive traffic” error.
I got nowhere trying to re-install the integration.
Has anyone tried reinstating the previous version of the HA OS (ie 11.4)? I’ve just checked that the OS update did not automatically create a backup that I can restore to and I’m not sure of the procedure to go back to the previous OS.
Edit 2hrs later. I’ve just reverted to OS 11.4 & I’ll post whether the error clears
On my side I deactivated rhe integration on HA and took at least 1h before I could use again the MELCloud app.
I’m too lazy (and mainly afraid to make it weong as I’m not at home) to rollback to previous version of HAOS.
But I’m suggesting you to start bu disabling rhe integration until you get again the access to MELCloud app, to ensure you unlocked t, then reactivate the integration, and confirmed it’s not blocked again.
Don’t delete the integration!
I deleted it, rebooted, and when I try to add mlcloud I get the information that with this credentials the integration exits and the key is changed, but no melcloud showing up!
The MELCloud integration is already configured for this email. The access token has been updated.
I waited 3 hrs after rolling back the OS but I am still getting the “excessive traffic” message. I’ve just disabled the integration to see whether that grants me access again. As I am also using REST sensors I fear it won’t.
I’ve ordered a bunch of D1 minis, will update you on this topic whether I’ll be able to manage my HVACs locally without melcloud
Ridiculous. I sent a snotty email to their support form. If I get anything useful back I’ll post it here. I suggest you all do the same.
The ESPHome solution sounds very tempting. No reliance on the cloud when it’s 2°C outside. It involves soldering, which is fine but my skills are not what they used to be, which was not much
Argh it’s so stupid! I’m sending just as many requests… And MELCloud replies! Flow and return temperature sensors are still updating, but I can’t make any changes
Hey guys,
I got a pretty reasonable reply:
Since the integration is basically not being managed by HA any more we’ll have to come up with a way to distribute the change so that those who want it can retain access.
For the moment I’ve updated my own local melcloud to set
MIN_TIME_BETWEEN_UPDATES = timedelta(minutes=5)
in __init__.py
and
SCAN_INTERVAL = timedelta(minutes=5)
in climate.py
(although I don’t think it uses that value…
I’ll report back when my access is restored.
Bruce
Just made those same changes after my access was restored - thanks for the suggestion.
I’ve also created 5 new “guest” accounts in the MELcloud app, using the “+” feature of gmail (i.e. i used [email protected]
) to register a bunch of different accounts to my same email address. Those new accounts had full control via the MELcloud app, even when my original account was locked out for excessive traffic. I’ll be using different accounts for automation from now on. Hopefully if this happens again, I will at least have some backup account that can control my heat pump manually…
Sorry if this is a dumb question, but are you guys using the Melcloud integration from HA or a custom component? I’m suffering from the same problem that my account gets locked because of excessive traffic. But I’m using the built-in component and there are no files to edit the scan interval, right?
I’m using the Melcloud integration from HA, not a custom component. I run HA in docker, so I had to do docker exec -it homeassistant bash
to get a shell, then edit the files above in /usr/src/homeassistant/homeassistant/components/melcloud
manually. This isn’t a permanent fix for me though as those files will get overwritten if I restart my docker container, update it, etc. But it’s good for now!
I’m running Home Assistant OS in a virtual machine and that seems to have a different file structure. Have to search if I can find the file.
Quick edit: Looking at Home Assistant over on Github, Frenck has reported yesterday that a fix with lower polling frequency will make it into the 2024.2 release. Let’s keep fingers crossed that this works.
Hallo,
ich hatte die gleichen Probleme mit meiner Klimaanlage.
Deshalb habe ich die Integrationen in HA gelöscht.
Habe 1 Tag gewartet und mich dann in der Melcloud App angemeldet.
Dort habe ich ein neues Passwort vergeben.
Jetzt konnte ich die Integrationen in HA wieder installieren.
Melcloud und Melcloud Custom.
Jetzt funtioniert so wie es scheint erstmal alles wieder.
Schauen wir mal wie lange.
Probleme fingen nach Update von Home Assistant an.
Version : 2024.1.6
This could be a dirty trick if the integration could allow entering several accounts using them to pull data on a sequential way, account by account.
Looks like a fix is coming today: MELCloud - Error - "This request has been throttled due to an excessive amount of traffic to our service" · Issue #109728 · home-assistant/core · GitHub
I haven’t tried but I guess it’s possible to use the pre-release of 2024.2 from here: Docker
I think the stable one is coming today anyway… but in case someone really can’t wait.
Man I wish a local protocol was mandatory… Zigbee… Matter… Something.
I think the reaction from Mitsubishi has been reasonable this far. Or even lenient. To my knowledge no accounts have been permanently disabled and the throttling periods aren’t event that long. The emphasis here is on the “so far” bit.
Now that they have intervened the message is pretty clear: the current level of data scraping is not something they want to handle.
It is definitely possible to work around the limitations by using an array of guest accounts or better MELCloud app emulation, but I can’t imagine that to work for too long. Either the throttling period gets longer, the threshold for throttling gets lower or accounts start to get permanently disabled.
IMO the load from a single account should not exceed that of a real MELCloud instance. There is a finite budget for providing a service for the lifetime of a device like an air-to-air heat pump and those numbers weren’t crunched with this load in mind.
Some guidelines I would follow:
- Only a single service should be polling data. Be that a rest sensor, nodered or a home assistant integration. If more services are a must, the polling rate should be adjusted to keep the load in the same ballpark.
- The polling interval should be at least 1 minute. Preferably much longer. Especially now to avoid further actions.
- Only the basic device info is polled that often. No report polling.
- Automations should not update values constantly. Heating/cooling is pretty slow anyways.
Please read the MELCloud EULA.