sidebar does not always automatically update when returning

[Mariusthvdb]

Having not been here for some time (...) please let me ask whether this is expected, or whether I should be doing something to fix.

Most of my sidebar items have a notification with the state of some sensor.

on the incidentally updating entities in that notification, the issue is not visible, but, when its a power sensor, the issue is immediately visible.

the same, or maybe even more obvious, is happening with my top title, which is showing the time.

What is the issue? returning after some period, be it 15 minute, or an hour or longer, the sidebar shows the lag in the time title,

partials:
  title: |
    const tijd = states(date_time);
    const tijdStr = new Date( tijd )
      .toLocaleTimeString([], {hour:"numeric",minute:"2-digit"});


title: |
  [[[
    @partial functies @partial title
    window.titleClick = titleClick;
    const event = 'ontouchstart' in window
      ? 'ontouchstart'
      : 'onclick';
    return `<span style="cursor:pointer"
              ${event}="javascript:titleClick()">
              Rumah kami: ${tijdStr}
            </span>`;
  ]]]

and in the power notification, one can see all sensor state changes fly by, until the current time has been reached....

  - item: energy
    match: href
#     order: 7
    icon: >
      [[[ @partial energy
          return nettov ? 'mdi:home-lightning-bolt'
                        : 'mdi:home-lightning-bolt-outline'; ]]]
    icon_color: >
      [[[ @partial energy
          return energy_color; ]]]
    notification: >
      [[[ @partial energy;
          return states(netto); ]]]

is there anything we can do to have the sidebar update immediately to its current state for all states in the templates?

we now need to wait, or manually reload/pull to refresh
thx for having a look

[elchininet]

Hi @Mariusthvdb,
Do you have the Suspend background connections switch disabled?

[Mariusthvdb]

Otoh I also get this in iOS app when opening.

[elchininet]

Ok, just test it and write down all the issues. Something to note is that before trying to update the status of an entity, one should ensure that the connection is open, this is something that I didn't take into account.

[Mariusthvdb]

I guess this is the most pressing error currently.
Cant open a browser without this happening. and because of it, cant check if the other updates are correct.

[elchininet]

I'll try the following:

1. Before calling the update_entity service, I'll check that the WebSocket connection is open, if it not open I'll wait for it.
2. If the update_entity call is rejected, I'll log the error to see if there are more details.
3. Right now I'm logging the name of the entity when the call is success. I'll do the same when the call ends in an error, so it would be easier to detect which entity failed.

With this I think that the testing will be easier. But right now I can explain what happens and why you are facing the issue:

During the idle state, if you disable the Home Assistant option to do the same, the connection is throttled or killed by the browser (this is something that browsers do to save resources, bandwidth, battery, etc). When one returns to the dashboard, the connection tries to reconnect and this happens pretty fast, but the state of the hass object (the one used for front-end applications) remains staled because it was not updated during the idle period. The hass object will be updated the next time that an entity updates. This is not a problem for entities that update regularly, but for entities like the the date_time_iso, that process will occur in the next minute and if you return to the dashboard after a minute started, that delay will be noticeable because it will take around a minute. I detected that calling the update_entity service, forces the update of an entity and refreshes its state in the hass object. This can help to solve the issue but I need to detect if it works well and in all the cases.

I'll let you know when I have a new bundle to be tested so you can do more debugging to detect possible issues with this approach.

[elchininet]

@Mariusthvdb,
Install the updated bundle in the gist and debug the behaviour with it. This is what you should expect in the logs:

1. Every time that you return to the dashboard after being hidden, it should log Updating... + the date/time of the update (in this way we can check when this event is triggered and how many times)
2. It should wait for the connection to be stablished, if the connection is alive it should continue, if not it should wait (maximum around five seconds).
3. If the connection is not ready in five seconds, then it should log The WebSocket connection was not ready. Cancelling the updates.
4. If the connection is ready then it proceeds, for each entity that was tracked before it will:
   1. Log Trying to update {id of the entity}
   2. If the update is success it will log {id of the entity} was updated, updating its templates and then the templates that used that entity are re-renderer
   3. If the update is rejected it will log {id of the entity} was NOT updated, printing the error and then a console.error with the content of the error

Let me know the results and then we can check what works and what fails.

[Mariusthvdb]

its very difficult to nail down, but even though this seems to be a lot better, I occasionally still get the toast piling up.

Whats more, I also see this. I believe per half hour:

Logger: homeassistant.components.template.coordinator
Bron: helpers/debounce.py:151
Integratie: Template (documentatie, problemen)
Eerst voorgekomen: 15:05:53 (1 gebeurtenis)
Laatst gelogd: 15:05:53

Unexpected exception from <bound method DataUpdateCoordinator._async_refresh of <homeassistant.components.template.coordinator.TriggerUpdateCoordinator object at 0x4d51287bce0>>
Traceback (most recent call last):
File "/usr/src/homeassistant/homeassistant/helpers/debounce.py", line 151, in _handle_timer_finish
await task
File "/usr/src/homeassistant/homeassistant/helpers/update_coordinator.py", line 435, in _async_refresh
self.data = await self._async_update_data()
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/src/homeassistant/homeassistant/helpers/update_coordinator.py", line 314, in _async_update_data
raise NotImplementedError("Update method not implemented")
NotImplementedError: Update method not implemented

could it be the bundle causes that? (as its practically the same error in the toast..)

ill revert to the release version next, and see if that error goes away

[Mariusthvdb]

yep, toast is gone after returning to the release version, and also the logger is silent now.

[elchininet]

The most important are the logs. I placed the logs to know what is failing and in which moment and hopefully to know why.

So if you can get some logs and let me know when they are triggered it would be helpful (I added the date/time in the first log, so you can check the time before going to the dashboard and know if the logs started when you returned or they ocurred during the idle moment)