My HA automation not always execute, especially for fast re-use. Analyze executed steps I found that some trigger condition based on HA entity from Z2M not pass (false condition check). However in real the condition should be true.
My question is can be above issue connected to how fast I write data to DB? My recorder setting is set to write every 60s but my trigger check condition of entity after 2-3sec after real change related entity.
So can be this true: Z2M device change own status and send MQTT update message to HA, HA receive it and process almost in real time. But HA entity will update DB up to 60s (let say average after 30s). Parallel after send Z2M message to HA I wait 2-3sec and execute trigger where is checked condition of the entity. What will be result? DB not updated but Z2M MQTT has been processed in HA?
What check automation entity condition, DB record of entity or real state from memory (if something that exist)?
I’m using Wyse 5070 as VM server, assign to HAOS 2 core, 16 GB RAM from 30 GB available, 120 GB GB (it is SSD Nvme), 1 GB LAN., I would not searching problems here.
Yes, my DB is huge 6.7 GB sqlite because keep data 30 days but general no facing issues performance, response is very good.
The device is local Zigbee. And yes, have a lot MQTT messages, maybe here is some delay in queue. Do you know how check MQTT queue, how long processing take place?
And on end DB, as I wrote I save DB every 60s. My question what is functional logic behind HA automation in my case? The automation checking entity state read from DB or from “memory”?