I am wondering about the step-by-step process of FP2 Presence Detection Automations, and how to get the fastest execution. Specifically, which automation triggers are event driven, and which are polling driven. FYI, I do not have an Aqara Hub, but I do have a Smartthings Hub and a Home Assistant server. I mostly use Home Assistant.
For Alexa automations, do FP2 presence events go to the Aqara Cloud, then to the Alexa Cloud, and then to my Action device? If so, how long do these steps typically take with a fast Internet connection?
For Smartthings, I see that my FP2 automations ran locally on my Smartthings Hub. But how does the Hub know when a Presence event occurs? Does if have to poll my FP2 sensor to discover the change in state, or does the FP2 sensor notify my Smartthings Hub immediately? If the change in state is detected via polling, how frequently does this happen and is there a way to poll more frequently?
For Home Assistant, my automations run locally on my HA server. But how does the server know when a Presence event occurs? Does if have to poll my FP2 sensors to discover any changes in state, or does the FP2 sensor notify my HA server immediately? If a change in state is detected via polling, is the a way to poll more frequently, and what would the specific steps be to accomplish this?
Nearly all my automations are done in Home Assistant and they work without an Internet connection. But I am wondering if Conditions such as an FP2 illuminance level adds delay to the execution time of the automation. I have some Presence Automations that trigger lights being turned On only if the brightness in the room is below a certain lumen level. (If lumen level conditions add delay, I would like to know so that I can store brightness levels as a local server variable that I update every few minutes rather than when a Presence automation is triggered.)
Basically, I want to know the fastest way to get a light turned On when Presence in a Zone occurs. I do use local actions (for example, a matter switch) to turn On lights ASAP, but are there methods that work faster than others? Currently, my lights turn On 1-2 seconds after my wife enters the room, which she notes is 1-2 seconds longer than if she had just flipped the light switch when she walked thru the door.
I suspect that my FP2 Presence automations are affected by a polling interval in Home Assistant. Is that correct? And if so, does a poll gather all the data at once (the presence state of each zone, plus luminance level)?
Thanks for any enlightenment you can provide me on making my lights respond faster to presence.
@tvlorsfamily Your inquiry into optimizing the speed of FP2 Presence Detection Automations is quite insightful! Let’s unpack your questions:
Alexa Automations:
Typically, FP2 presence events are sent to the Aqara Cloud before reaching the Alexa Cloud and then your action device. This multi-cloud interaction can introduce latency, especially reliant on your internet speed. However, specific timing details aren’t provided in the reference, but a fast internet connection should minimize delays.
SmartThings Hub:
Good news! The FP2 sensor is designed to notify your SmartThings Hub immediately when a presence event occurs, rather than relying on polling. This should inherently make the process faster as real-time notifications are generally quicker than polling intervals.
Home Assistant:
Similar to SmartThings, the FP2 should notify your Home Assistant server immediately when a presence event occurs. If you’re experiencing delays, it might not be related to polling intervals. Unfortunately, detailed steps for altering polling frequencies are not available in the reference data. However, ensuring that your Home Assistant setup is optimized for immediate event notifications could help reduce latency.
Regarding the illuminance level conditions, these might indeed introduce some delay due to additional processing needed to check the condition against the current light level. Storing brightness levels as a local server variable and updating them periodically, as you suggested, might help streamline the process.
To achieve the fastest light activation, using local actions like a matter switch is a good approach. Ensuring that all automations run locally and minimizing cloud dependencies will generally yield quicker results.
Lastly, while polling could influence automation speed, the FP2’s design to send immediate notifications suggests other factors might be contributing to the delay you’re experiencing. Checking your local network setup and ensuring all devices are on the latest firmware could also help.
I hope this helps speed up your automations! Feel free to share more insights or questions as you continue to optimize your setup.
For clarity, here’s how FP2 Presence automations actually work in different platforms, and how to minimize latency:
Alexa
Yes, presence events from the FP2 go through the Aqara Cloud, then to the Alexa Cloud, and finally to your action device. Each cloud hop adds delay, usually ~500 ms per hop. With a fast Internet connection, total latency is normally 1–2 seconds. That’s why cloud-based automations feel slower.
SmartThings
SmartThings can run automations locally on the hub. The FP2 does not require polling; it pushes presence events immediately to SmartThings via its integration. So the hub knows about state changes in real time. No manual polling interval is involved.
Home Assistant
Same principle: FP2 reports state via push notifications. Home Assistant does not poll every few seconds; the sensor notifies HA immediately when presence or luminance changes. So, no, there is no hidden polling delay you can “speed up.”
Conditions (like illuminance check)
Adding a condition like “only if lux < X” does not introduce significant delay. HA already has the latest illuminance value from the last push update. Evaluating that condition is basically instant (<1 ms). No need to move lux into a separate variable unless you want to optimize logic for readability.
Fastest way to turn lights on
Use local automations (as you do in HA).
In practice, FP2 → HA → Matter switch is usually ~500 ms to 1 s. The extra second you notice can come from the FP2’s detection algorithm itself (it waits to confirm presence before firing). That’s by design to reduce false positives.
Polling?
No. Neither SmartThings nor HA relies on polling the FP2 for presence or lux. It’s all event-driven. When FP2 sends an update, all states (zone presence or lux) are included in that payload.