[hue] Bridge (v1) going OFFLINE and then ONLINE frequently due to connection lost about openhab-addons HOT 14 OPEN

The real question is what is causing the connection to be lost?

Yes, that is the question,. I will add logs to the binding to understand in which case it happens.
Maybe due to authentication....

from openhab-addons.

In fact, there is a lock mechanism so concurrent requests when polling data is not possible.

from openhab-addons.

Here is the line triggering the update of thing to OFFLINE:

from openhab-addons.

I guess it is normal for the bridge to indicate offline if the connection is lost.

The real question is what is causing the connection to be lost?

from openhab-addons.

There was a fairly major firmware update for v2 bridges on Aptil 25th, so maybe they pushed something for v1 at the same time..

https://www.philips-hue.com/en-gb/support/release-notes/bridge

from openhab-addons.

After adding/updating some logs, looks like a timeout problem when polling the bridge:

2024-06-08 11:49:47.423 [WARN ] [ue.internal.handler.HueBridgeHandler] - Connection to Hue Bridge 192.168.xxx.xxx lost: Total timeout 5000 ms elapsed
org.openhab.core.i18n.CommunicationException: Total timeout 5000 ms elapsed
        at org.openhab.binding.hue.internal.connection.HueBridge.doNetwork(HueBridge.java:1128) ~[bundleFile:?]
        at org.openhab.binding.hue.internal.connection.HueBridge.doNetwork(HueBridge.java:1094) ~[bundleFile:?]
        at org.openhab.binding.hue.internal.connection.HueBridge.get(HueBridge.java:1077) ~[bundleFile:?]
        at org.openhab.binding.hue.internal.connection.HueBridge.getGroups(HueBridge.java:491) ~[bundleFile:?]
        at org.openhab.binding.hue.internal.handler.HueBridgeHandler$2.updateGroups(HueBridgeHandler.java:312) ~[bundleFile:?]
        at org.openhab.binding.hue.internal.handler.HueBridgeHandler$2.doConnectedRun(HueBridgeHandler.java:254) ~[bundleFile:?]
        at org.openhab.binding.hue.internal.handler.HueBridgeHandler$PollingRunnable.run(HueBridgeHandler.java:138) [bundleFile:?]
        at java.util.concurrent.Executors$RunnableAdapter.call(Unknown Source) [?:?]
        at java.util.concurrent.FutureTask.runAndReset(Unknown Source) [?:?]
        at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(Unknown Source) [?:?]
        at java.util.concurrent.ThreadPoolExecutor.runWorker(Unknown Source) [?:?]
        at java.util.concurrent.ThreadPoolExecutor$Worker.run(Unknown Source) [?:?]
        at java.lang.Thread.run(Unknown Source) [?:?]

Maybe 5s is a little short as timeout ? I am going to try with a bigger timeout.

from openhab-addons.

Maybe 5s is a little short as timeout ?

Yes this could indeed be the case. For example: if the Hue Bridge is running a Hue Labs automation such as 'Color Loop' then it responds much slower to HTTP requests.

Note however that the above problem may go away rather soon because Philips / Signify will remove Hue Labs from the bridges at the end of this month (if I recall correctly), and will have replaced it with newer built in Effects such as 'Prism' and/or other built in functions which perform faster.

from openhab-addons.

@andrewfg I am using a bridge v1 so I don't think there will be any change from Philips / Signify in the future.

Trying with 8s as timeout for requests, I have not yet encountered any problem. Lets see if it is reliable after few hours.

from openhab-addons.

It just happened again with 8s timeout.

from openhab-addons.

@lolodomo I suppose that being a v1 hub, you are using API v1 rather than v2 .. if so, I am pretty sure there was nothing recently changed on the OH side that could have suddenly started your issue after it working for so long. However it seems also unlikely that Philips / Signify has changed the v1 hub firmware. The only thing that I can think of may be the lamp firmware, but this seems unlikely to cause an HTTP comms issue..

• https://www.philips-hue.com/en-gb/support/release-notes/bridge
• https://www.philips-hue.com/en-gb/support/release-notes/lamps

from openhab-addons.

@lolodomo by the way, the API v2 specification states that the Bridge is limited on number of serial HTTP calls per second (as I recall about 20 per second), and number of parallel overlapping HTTP calls (as I recall about 2 or 3). And I suppose that a similar limitation would apply to API v1 resp. the v1 bridge. When I wrote the API v2 code I introduced a throttler to reduce HTTP calls to within the given serial/parallel limits. I did not back-port it to API v1 since I assumed if that if anyone would ever encounter HTTP limitations they could solve it by upgrading to API v2. Yet for your special case you could consider back-porting the throttler -- just a thought..

from openhab-addons.

Problem is unchanged with 10s timeput.

As I have the sensor polling interval set to 0, I should not have concurrent requests, except of course if all requests run when polling lights update sometimes takes more than 10s. I mean maybe the next cycle starts while the previous one is not ended.
In case it is the problem, I could try to increase the refresh interval to see if it helps.

from openhab-addons.

One solution or rather a workaround could be to not set the bridge thing to OFFLINE when there is a request timeout during the data polling but rather do it if it happens during two consecutive polls.

from openhab-addons.

I am testing a basic retry mechanism added for GET requests failing in timeout and it seems to work well.
Will see if it is fully reliable after several hours.

from openhab-addons.