Subscription failed permanently

I have a remote daylight monitoring system that has been running for over a year. I have two devices at home that monitor a published event from this system once every 60 seconds during the day and once every 15 minutes at night.

Over the last 2 weeks I have noticed that my two devices at home have been losing their subscription randomly. A reset would bring them back for a couple of days and then they would lose their subscription again. I changed the code on one to automatically reset when this happened and that worked for a few days. Two days ago at 3.35pm my time both these devices stopped listening to the published event and no matter what I do they will not respond to the published event again.

The oldest one was running firmware 0.4.7 and has since been upgraded to 0.5.1 in a vain attempt to see if that made a difference, no success. Subscribing to the published event returns with success

The second device was running 0.6.0 and is now running 0.6.1. It too says it is subscribing successfully but does not respond to the published event either.

I know the event is being published as I can see it on the console and other services like IFTTT are still gathering data based on the event. It is published as a public event and both of my Photons at home subscribe to this event using the publishing photons ID. I am not using the MY_DEVICES option.

Both my devices at home are Photons and are permanently powered. The remote device is also a Photon and sleeps for the better part of each cycle.

I have no idea why something that has been working on it’s own for over a year is suddenly failing. Any help to guide me to poke the right thing is much appreciated.

Thanks in advance,
Joerg Seiler

@Jseiler, is the Photon listening to events from another Photon of yours?

Yes it is.

The remote Photon is self subscribing to it’s own event to ensure the event got published. One of my Photons at home listens to this same event and publishes a second event that the remote Photon listens to to ensure a round trip.

My other Photon at home listens to only the event published by the remote Photon and does no publishing of it’s own.

Alright and which part(s) is/are failing?
Does the remote Photon’s self subscription fail or?

What are you using to determine that the subscription is failing?

Sorry but i wanted to get as much context as possible and see if anyone can on the logs.

I don’t have immediate access to the remote Photon so i can’t tell if it’s failing it’s self subscription. It is however failing the round trip part of it’s code as my home unit is not responding to it.

The two Photons I have here beside me are logging the data and keeping time based on the remote Photons event. I can tell they are not hearing the event as they are not actively updating their displays or logging data.

In the mean time I have flashed my home code on another Photon I have and it is also not hearing the event although it is subscribing successfully as well. So this is not device related.

Can you let me know the event name here? Otherwise, you can drop me a PM.

Not a problem. The event name is Daylight and it is published by a Photon named Ocelot

I am running into the same issue, I believe. I’ve been publishing an event from IFTTT that’s triggered from my Amazon Echo to turn on my lights. Suddenly, my particle photon is not hearing the event. I can see that IFTTT is publishing the event (or trying to at least), and I know that my particle photon is online.

I should add, that the whole setup has been working fine for months. The issue has come on suddenly.

Hello folks,

Indeed I did a test and there’s some weirdness with the public firehose subscription. The team is aware and cloudy folks are going investigate!

1 Like

Cool. Thanks for the quick investigation.

Status update:

Investigating
We’ve received multiple reports of devices not receiving all the expected events and are investigating.
Mar 22, 20:16 PDT

New Incident Status: Identified
We have identified the issue. It only affects TCP devices, and we are implementing a fix now.
Mar 22, 20:59 PDT

New Incident Status: Monitoring
We have deployed a fix and are monitoring. Events appear to be being received as expected.
Mar 22, 21:25 PDT

@Jseiler,

Things should be back to normal. Let us know how it goes! :slight_smile:

Checked first thing this morning and am pleased to report that the situation seems to be back to normal.

Thank you so much for jumping on this so quickly. :relaxed:

3 Likes

I can confirm that my device is now operating normally as well. Thanks.

1 Like