pricemi115
commented
1 day ago Hi @ScottBessel Thank you for the report. I am guessing that the packet loss issue is a similar issue as #61. I can look into that and have a look.
The names of the 3 accessories under the single device has always been a challenge. It seems that homebridge does not percolate these names very well. But it has been a while since I looked into it, so I can look to see if there are any improvements there.
When I created this plugin, my thought process was to use the CO sensor as that reports "bad stuff" and the when ping results get large, that is also bad. But I like the idea of using other accessory types and sensors, like motion sensors, which would support automation.
Out of curiosity, what kind of automation would you want to do is the results got into the bad zone?
ScottBessel
I just installed version 1.6.1 of this plugin on a raspberry PI. I am not sure where to put this issue, but since this was where I was directed, here it is.
OS Raspbian GNU/Linux Bullseye (11) Homebridge version: v1.7.0 UI version: v4.56.4
No matter what I do the packet loss always shows zero. If I set the alert to packet loss I get a lot of random alerts, all I want is to be alerted if it is actually down, hence the high loss_limit number, I tried this on a variety of numbers, they seemed to be ignored. I am not sure what is tripping the alerts, as when I get the packet loss alerts there is no packet loss and the connection is still good. If I turn the alerts all off, as it is set now in the config, I still get alerts for latency, even though the latency is BELOW the setting I have of 200.
On a side note: you labels in the UI are not the same as the config file, as such the documentation does not correctly relate, so the loss_limit is a confusing setting.
Feature: It would be nice if I could pick the device type, I.E. light, switch, motion sensor. The carbon dioxide detector has a particular nasty alert and does not offer as many automation options as the others. The carbon dioxide alerts don't allow narrowing of the device, just the room, as they are all in the same room the alerts are pretty much meaningless until you drill down to the actual device. A light switch that I can label "Google Packet Loss" turning on would be simple to see and trigger on.
here is the config: { "settings": { "history_logging": { "reporting_period": 10, "maximum_history_size": 250000, "retention_period": 30 }, "ping_targets": [ { "target_type": "ipv4", "modem_type": "other", "target_dest": "8.8.8.8", "loss_limit": 90, "expected_latency": 200, "expected_jitter": 20, "peak_expiration": 12, "data_filter_time_window": 180, "sensor_alert_mask": 0 }, { "target_type": "uri", "modem_type": "other", "target_dest": "google-public-dns-b.google.com", "loss_limit": 90, "expected_latency": 200, "expected_jitter": 20, "peak_expiration": 12, "data_filter_time_window": 180, "sensor_alert_mask": 0 } ], "ping_period": 20, "ping_interval": 1, "ping_count": 5, "packet_size": 56, "enable_history_logging": true }, "_bridge": { "username": "0E:65:DF:EB:86:E8", "port": 33497 }, "platform": "GrumpTechHomebridgeNetNanny" }