No feedback on registration for IoT Rules Engine


#1

Since Particle IoT Rules Engine was announced, I subscribed to the beta. I never got any feedback (yes or no), so I tried again 2 times, to no avail.
Is there any particular restriction on this fonctionnality at the moment? Like something is not working, and it’s not Particle’s priority at the moment?

I am just a hobbyist, I am not building a product. But still I’d like to know if I should wait or not.


#2

There is a long waiting list for the rules engine beta. Later in 2019 there will be a public release of the rules engine, which will be open to everyone. Right now, mostly product creators are in the beta program.


#3

I can understand your choice to open the beat only for creators, even tough it’s frustrating :slight_smile:
That said, I don’t remember reading anything about the beta beeing closed to hobbyists at the moment. Could be good to communicate on that so users are not left in the dark.


#4

That’s not what @rickkas7 said. It’s not limited to any particular group but the number of available test accounts was/is limited and they are already filled.
Whether and when new spots open is not clear, but as soon they’d open they’d probably filled instantaneously with peoplo on the waiting list as first-come-first-serve.

The statement about who already is in the program was a side note, not a statement of limitation.


#5

@rickkas7 We are almost at the end of 2019, more than 1 year after the Spectra announcement. I have the same issue, no feedback on the inscription of the beta program. You wrote there will be a public release at the end of 2019. When will it be released? You can keep people waiting but this is really too long before the announcement and the release.


#6

The beta program for the Rules Engine has ended. It should not have been possible to sign up recently, and if you had previously signed up you should have received an email. In any case, the rules engine as demoed at Spectra won’t be released as a product as-is. At some point in the future, there may be an updated rules engine feature, which will have a new beta program, hopefully leading to an actual release.


Particle.function node - Rule engine
#7

@rickkas7 thank you for your feedback. This is very unfortunate. Maybe you should make a statement in your website then as well, I think there are a lot of people like me who wanted to use this (the easy of Node-RED is awesome) in their product development. It is still listed in your products list on your website (even as NEW - BETA).

So this is completely dead? Or still ives upon request based for product owners?


#8

Hey there,

Want to clarify current positioning of the Rules Engine instance and address concerns raised in this thread.

As Rick mentioned, interest in the Rules Engine has been very high. Throughout the course of the beta, we’ve opened access to the Rules Engine to a variety of Particle customers, including hobbyists, enterprises, maker spaces, universities, and corporate R&D labs.

We’ve learned how Rules Engine can help customers with their IoT deployments, and gained lots of experience working with Node-RED, the technology that powers the Rules Engine.

One of the primary challenges of Node-RED as an underlying technology is that it was originally designed to be deployed as a monolithic application, designed to be run on devices like a Raspberry Pi. This has created two problems during the beta:

  1. The editor and runtime are linked. When the Rules Engine instance crashes, Particle (the infrastructure provider) needs to manually intervene on behalf of the customer. This behavior has been improved with the recent v1.0 release of Node-RED.

  2. Node-RED cannot easily be deployed in its present form in as a multi-tenant service, meaning that every instance requires its own discrete infrastructure that requires uptime instrumentation and management. This means that the cost of maintaining instances scales roughly linearly with the number of users, which creates unique challenges for the product.

As a result of those scaling challenges, we limited further access to the beta. The product is not dead, but we are actively reconsidering how we deploy the product as a feature of the Particle platform.

Hope that helps to clarify things. More details to come.


#9

@will @rickkas7 Thank you for your clear response. This clarifies a lot. I understand the problems you are facing very well.

We in our company are actually facing similar issues. Right now we deploy them as a microservice (docker container running on AWS). But this is a container for each customer at this point in time (same as you are facing). We are in the process of setting-up an automated pipeline for making this happen for each customer.

For the crashes, … depending on the cause this can be automated as well. We consider Node-RED just as any other microservice so we have a readiness, liveliness probe in our Node-RED installed, together with a metrics probe which are all monitored by CloudWatch. And CloudWatch is responsible for staring a job to kill and redeploy the container then. Even in Node-RED for some reconnecting purposes (after the retry counter has reached the end for a STOMP queue, we internally from our Node-RED call the Node-RED admin API to redeploy the flows automatically).

So I understand what you wrote. However, I think you should just have spread the word faster. People will understand but if you do not tell them anything, they get annoyed (just as I did because I was waiting for it just as I am for the production model of the 3Gen devices on building a product).

That all said, thank you for the update.

If at any point in time you want to discuss our experiences with this point (as you have read we have a lot of experience with it ourselves) just give me a sign and I will be glad to schedule a telco with my colleagues and you guys


#10

Thanks for the update @will.

I’ve also been waiting on this, and tried to sign up for the beta. I’d like to throw a suggestion that I think a lot of people would get value from even the start of an implementation.

For example for myself I’m forwarding events to io.adafruit and then doing very simple triggers that fire if a value ends up too low. The only options are to email me, call a webhook, or publish a new message. It’s enough though.