Climate offline, temperature sensor still reports value

[rr326]

The problem

I have some thermostat devices - some ecobees and a Nuheat. They create climate.* and sensor.xxx_temperature entities.

When the thermostat is offline, the climate will properly show state==None. But the temperature sensor will continue to show the last valid value - which is definitely no longer valid. This means I'm relying on a value that is incorrect. Further, I can't, easily, just use the "current temperature" value of the climate entity in my use case. (In my case, I use a custom card that doesn't handle null values well.)

I think the proper behavior should be when offline, the sensor should read None, or maybe numpy.Nan

Environment

• Home Assistant Core release with the issue: 0.188.4
• Last working Home Assistant Core release (if known):
• Operating environment (OS/Container/Supervised/Core): Docker container, armv71, 5.4.72-v71+
• Integration causing this issue: ecobee
• Link to integration documentation on our website:

Problem-relevant configuration.yaml

# Configure a default setup of Home Assistant (frontend, api, etc)
default_config:

# Text to speech
tts:
  - platform: google_translate

group: !include groups.yaml
script: !include scripts.yaml
scene: !include scenes.yaml
automation: !include_dir_merge_list automations

homeassistant:
  name: Home
  latitude: !secret latitude
  longitude: !secret longitude
  elevation: !secret elevation
  unit_system: imperial
  temperature_unit: F
  time_zone: America/Los_Angeles
  internal_url: http://localhost:8123
  customize: !include customize.yaml
  packages:
    package_inputs: !include package_inputs.yaml

lovelace:
  mode: yaml
  resources:
    - url: /local/button-card.js?v=1
      type: module
    - url: /local/simple-thermostat.js?v=1
      type: module
    - url: /local/banner-card.js?v=1
      type: module
    - url: /local/card-mod.js?v=1
      type: module
    - url: /local/bignumber_fork/bignumber-card.js?v=1
      type: module
    - url: /local/stack-in-card.js?v=1
      type: module
    - url: /local/auto-entities.js?v=1
      type: module

  dashboards:
    dashboards-all:
      mode: yaml
      title: Dashboards
      # icon: mdi:script
      show_in_sidebar: true
      filename: dashboards/main_dashboard.yaml

zwave:
  network_key: !secret zwave_network_key
  
# Turn on http: needed by api
http:
# Turn on api: needed by MQtt
api:

logger:
  default: info

mqtt:
  broker: 127.0.0.1
  port: 1883

notify:
  - name: event_log
    platform: file
    filename: /config/event_log.log

  # See DEVNOTES.md for details
  - name: gmail
    platform: smtp
    server: smtp.gmail.com
    port: 587
    encryption: starttls
    timeout: 30
    sender: !secret gmail_un
    username: !secret gmail_un
    password: !secret gmail_app_password
    recipient: !secret email_recipient
    sender_name: !secret email_sender_name

Traceback/Error logs

Additional information

If this is indeed a bug (and not a side effect of something needed or a design decision), if you point me a the likely solution, I'll try to fix and issue a PR.

[frenck]

So which integration is this bug report for? We can only target one integration per issue...

[rr326]

I'd start with the ecobee integration. I know that is a problem. But I'm pretty sure it is in the NuHeat integration as well. So either both integrations copied the same code (and bug), or could it take place outside the integration?

[frenck]

No, this behavior is defined by the integration itself.

[probot-home-assistant]

ecobee documentation
ecobee source
_{^{(message by IssueLinks)}}

[probot-home-assistant]

Hey there @marthoc, mind taking a look at this issue as its been labeled with an integration (ecobee) you are listed as a codeowner for? Thanks!
_{^{(message by CodeOwnersMention)}}

[hakusaro]

There's probably a lot of good value in at least publishing an announcement about how users should re-add their integrations. This is my first real experience with HA, and it seems like some parts of the stack don't report errors back in a noticeable way, particularly when it comes to automations. The only way I knew it was as simple as re-adding the integration was finding this thread. Granted, the ecobee API is not intuitive and should probably offer some scheme for users to be able to continue to use existing apps -- but the UX here is pretty meh given that the underlying everything still technically works, but you need to do work to re-add it for existing setups to continue working.

Is it possible to add something like this to release notes / changelogs / etc easily?

[rr326]

@marthoc I've been looking in the code at components/ecobee/sensor.py. It's all new to me, but I think the logic is:

• Wait for an update from the EcobeeData object
• If the capability is 'temperature' (or 'humidity'), then update the state
• Otherwise ignore and move on

This would definitely cause the problem I'm seeing since when the ecobee is offline, it doesn't have a 'temperature' capability.

The relevant code is here:

    async def async_update(self):
        """Get the latest state of the sensor."""
        await self.data.update()
        for sensor in self.data.ecobee.get_remote_sensors(self.index):
            if sensor["name"] != self.sensor_name:
                continue
            for item in sensor["capability"]:
                if item["type"] != self.type: # <-- If offline, there will be no 'temperature' / self.type
                    continue
                self._state = item["value"]
                break

You'd probably need to add a:

if self.type not in sensor["capability"]:
   # Offline
   self._state = math.nan # Can't set state to None

It looks to me like this is generic code. It probably affects all the sensors in ecobee, and quite likely other components that copied this logic.

If that's the case, this could be a big fix across many components that would need to be thought through. Definitely above my pay grade for a PR!

[github-actions]

There hasn't been any activity on this issue recently. Due to the high number of incoming GitHub notifications, we have to clean some of the old issues, as many of them have already been resolved with the latest updates.
Please make sure to update to the latest Home Assistant version and check if that solves the issue. Let us know if that works for you by adding a comment 👍
This issue has now been marked as stale and will be closed if no further activity occurs. Thank you for your contributions.

[rr326]

@frenck (or someone else) - I think someone with good overview of the HA system and feature / bug roadmap should look into this. I do not need it for me, but if I'm right, then many or most integrations are publishing meaningfully incorrect data. It seems a pretty big and fundamental error, if I'm right. Someone should look into and either disprove, or put in the appropriate place on the roadmap.

[github-actions]

There hasn't been any activity on this issue recently. Due to the high number of incoming GitHub notifications, we have to clean some of the old issues, as many of them have already been resolved with the latest updates.
Please make sure to update to the latest Home Assistant version and check if that solves the issue. Let us know if that works for you by adding a comment 👍
This issue has now been marked as stale and will be closed if no further activity occurs. Thank you for your contributions.