Upcoming SmartThings API Usage Limits for SharpTools Rules

[josh]

I hear you guys and I’ve been feeling the frustration and stress of this as well, so I can empathize!

@Nezmo

I’m already in timeout .

Enforcement is scheduled to start October 1, so you should just be seeing the warning, but no actual limitations.

I wanted to get the notices out as early as a reasonably could, but there’s still some rough edges. Lots of work to get to this point and with the timeline from SmartThings, I still feel like I’m pushing things out a bit earlier than ideal to try to give everyone enough time to not be caught off guard last minute. (more on that in my reply to Nathan below)

@josh

…made significant improvements behind the scenes to reduce the number of SmartThings API calls SharpTools Rules need to make.

I’ve been working hard in the background to reduce SmartThings API calls as much as possible. A bunch of work around enabling more detailed attribution, and metrics around API calls. Then analyzing and finding areas of opportunity for reduction. A bunch of work behind the scenes to seamlessly transition conditions and device state reads in rules to set up event subscriptions so we could read that state from the local database instead of API calls.

But there are still fundamentally some patterns that cause higher usage and impact the top few percent of SmartThings users.

To that point, this specifically impacts SmartThings API calls based on the change of SmartThings is rolling out. Other integrations like Homey, Home assistant, and Hubitat are not impacted.

I also built a local SharpTools Bridge with a bunch of LAN integrations like Phillips Hue, TP Link Kasa, Lutron, Fully Kiosk, and Matter.

SharpTools Bridge private alpha is starting

Using one of those alternatives that isn’t directly impacted could be useful for some cases.

---

@Nathan_Maxwell

How can I put my rulebase on an API diet when you’re lumping almost all the calls into an “other and unattributed” category? Should I just hunt and peck through all 57 rules and hope I find the culprit?

I should I have clarified how to use the Usage reporting and the current limitations. The warning you see in amber is a key point:

This time range includes days without rule level tracking. Treat it as directional.

The overall usage shown in the top bar has been in place since before the beginning of the month.

The attribution per rule was only added August 14 which is why you see the amber warning on the month range selection. September should be the first month with full rule level usage tracking where that amber warning will disappear.

In the meantime, if you filter to the latest day, you should get more accurate rule level tracking.

And similarly, once August 23 rolls around, the seven day usage should be complete.

—

My apologies as I know it’s not ideal. I felt it was important to get out whatever metrics we had available as early as possible, even with the rule level attribution being incomplete at the monthly level.

—

From a quick review, a few opportunities:

1. House Off Fanout: One execution fans out across roughly 60 devices. Moving the bulk shutdown into a SmartThings Scene (Manual Routine) and executing that Scene from SharpTools could turn dozens of SharpTools API calls into one.

2. Motion Rules: Lights On @ Active repeatedly resend commands to groups of devices on motion events. State guards could prevent commands when the lights are already in the desired state. A SmartThings Scene could also reduce each multi-device execution to one SharpTools API call.

3. Fade Rules: long sequential fade macros generate repeated level commands. Conditional usage, sparingly using these, or single level transitions would reduce usage.