Just a quick update as I have it working but I have new question 
Table now looks like this. Just need to work out how to get the RSSI for each device but this is not part of connected_devices entity. Any ideas?
I may have encountered a significant performance issue with this integration. Iâve been experiencing high load recently and I finally ran the profiler with the following results:
After disabling the integration, the results were the same (indicating this integration may not clean up itâs resources properly when disabled). After a restart my profiler results were better:
This pretty clearly indicates to me this integration was causing the performance issues. My system load also dramatically decreased after the restart.
Iâm posting this here as a first step and to get quicker/wider feedback. Iâll also be submitting an issue on github with more detailed information.
I hope you dont expect ânormalâ people should understand your posted pics and ânumbersâ
In short i understand " significant, and doesnât clean up " , and i also have been searching for âpowerâ consumers in my setup, as it has increased , over the past 3-5 month, you could say âsignificantâ but as i run on a I5 laptop, im more talking about increase in CPU usage 1-2%, but still itâs about 50% increase without adding devices, automation etc. ⌠how much the monthly HA updates causes of this i dont know, so i welcome your research, if it can bring it back down ⌠in the long run IO and CPU usage cost money, and wear on disk
No, I donât expect non-technical people to understand profiler diagrams. Iâm posting it here for exposure and for anyone who is able to understand and comment on it to do so.
So what are we talking about, in terms of CPU, IO and Memory consumption ? , or are you also referring to response-time ?
I do however know that in particular 1 update( think it was this April ), slightly increased cpu and memory, in favor of less disk-writes
And i did âfollowâ your approach, disabled this integration, rebooted the host , slightly noticeable , in either cpu usage or memory consumption, not even " a pin " in Power-consumption
Thou ofcause, i have no idea how you use this integration
Hi all, I have this integration working quite well now - thanks to many contributions on this thread and others! Dashboard looks like thisâŚ
Really keen to get some assistance in getting RSSI included in the connected devices table but I canât find anything on thisâŚ
Code if anyone is interestedâŚ
type: vertical-stack
cards:
- type: custom:mushroom-chips-card
chips:
- type: template
entity: sensor.rt_ax86u_connected_devices
content: >-
Wired: {{ state_attr('sensor.rt_ax86u_connected_devices',
'devices')|selectattr('connection_type', '==', 'Wired') |list|count }}
icon: mdi:lan-connect
icon_color: green
double_tap_action:
action: none
tap_action:
action: none
hold_action:
action: none
- type: template
entity: sensor.rt_ax86u_connected_devices
icon: mdi:wifi-strength-2
content: >-
2.4 GHz: {{ state_attr('sensor.rt_ax86u_connected_devices',
'devices')|selectattr('connection_type', '==', '2.4
GHz')|selectattr('guest', '==', false) |list|count }}
icon_color: blue
double_tap_action:
action: none
tap_action:
action: none
hold_action:
action: none
- type: template
entity: sensor.rt_ax86u_connected_devices
icon: mdi:wifi
content: >-
5 GHz: {{ state_attr('sensor.rt_ax86u_connected_devices',
'devices')|selectattr('connection_type', '==', '5
GHz')|selectattr('guest', '==', false) |list|count }}
icon_color: purple
double_tap_action:
action: none
tap_action:
action: none
hold_action:
action: none
alignment: center
- type: horizontal-stack
cards:
- type: custom:mushroom-entity-card
entity: button.rt_ax86u_restart_wireless
name: Restart WIFI
primary_info: name
hold_action:
action: toggle
tap_action:
action: more-info
double_tap_action:
action: toggle
fill_container: false
icon_color: deep-purple
secondary_info: none
- type: custom:mushroom-entity-card
entity: button.rt_ax86u_reboot
name: Reboot Router
primary_info: name
hold_action:
action: toggle
tap_action:
action: more-info
double_tap_action:
action: toggle
fill_container: false
icon_color: orange
secondary_info: none
- type: horizontal-stack
cards:
- type: custom:mushroom-update-card
entity: update.rt_ax86u_firmware_update
name: Firmware
- type: custom:mushroom-template-card
secondary: 'IP: {{state_attr(entity, "ip")}}'
primary: |
{% if is_state(entity, "on") %}
Internet Connected
{% else %}
Internet Down
{% endif %}
icon_color: |
{% if is_state(entity, "on") %}
blue
{% else %}
red
{% endif %}
icon: |
{% if is_state(entity, "on") %}
mdi:check-network-outline
{% else %}
mdi:network-off-outline
{% endif %}
entity: binary_sensor.rt_ax86u_wan
tap_action:
action: more-info
multiline_secondary: false
- type: custom:mushroom-entity-card
entity: sensor.rt_ax86u_boot_time
name: Router Last Rebooted (Uptime)
icon_color: brown
- type: horizontal-stack
cards:
- type: custom:mini-graph-card
name: Download
entities:
- sensor.rt_ax86u_wan_download_speed
- type: custom:mini-graph-card
name: Upload
entities:
- sensor.rt_ax86u_wan_upload_speed
- type: horizontal-stack
cards:
- type: gauge
entity: sensor.rt_ax86u_ram
name: RAM
needle: false
- type: gauge
entity: sensor.rt_ax86u_cpu
name: CPU
- type: custom:fold-entity-row
padding: 0
head:
label: Connected Devices
type: section
tap_action:
action: none
entities:
- type: custom:flex-table-card
sort_by:
- devices+
strict: true
entities:
include: sensor.rt_ax86u_connected_devices
columns:
- data: devices
name: ' Device Name'
modify: x.name
- data: devices
name: ' IP'
modify: x.ip
hidden: true
- data: devices
name: ' Connection'
modify: x.connection_type
- data: devices
name: ' Node'
modify: >-
if(x.node == "a8:5e:45:x:x"){"Balcony"} else if(x.node ==
"7c:10:c9:ec:x:x"){"Living"} else if(x.node ==
"3c:7c:3f:44:x:x"){"Bedroom"} else{x.node}
I am not getting tracker updates from this integration. Been that way for awhile. Devices never change to Away unless I reload the integration.
Any solution for this?
Does toggling the LED entity write to NVRAM or does it only change the value in the NVRAM mirror? I have merlin so I canât test it by rebooting.
Currently I have a cron job set up on the router to toggle the LED at night and in the morning, but I only write to the NVRAM mirror to avoid wear on the flash. But if this service doesnât write to the NVRAM either then it would be better to set up an automation in HA.
Hello, @szerencl,
The integration uses the FW service ctrl_led the same way as the Web UI is doing (via this button in the AiMesh panel).
As far as I know, it writes the value to NVRAM directly and on the stock FW the LED state is staying the same even after reboot. At the same time, Merlin ignores this NVRAM value and does not load it on boot (because long ago Merlin implemented a different way to disable LEDs). To work around it, when the communication library notices a reboot and if the last state of the LED is âoffâ, it actually sends to Merlin (and only Merlin) an âonâ/âoffâ sequence for the LED to restore the correct state. With the âonâ value from before the reboot, nothing is sent.
Considering the fact that this is exactly the same behaviour as the Web UI has, I donât think this should be a problem for the NVRAM wear. Especially considering the number of changes written to NVRAM constantly in any case. But this is up to you to decide what is better in your case.
@kramttocs, hello. I am constantly trying to improve the device tracking features. Unfortunately, the lack of any testing devices (except for the only RT-AX88U I have) and the fact that the issue might happen and not happen to the two identical devices with the same FW version is not helping in solving it fast.
But it will be solved eventually.
Can you please open an issue on GitHub with all the details about the problems you are experiencing, so that we can troubleshoot and solve everything properly.
To everyone else. If you are experiencing some issues and there is no such issue on GitHub yet, please open an issue. If the same issue already exists, please just comment on it with the device model and the FW version you are using.
With time itâs getting harder to track all the issues here and this can help with providing fixes much faster.
The new version is coming soon to AsusRouter. Please, stay tuned 
RT-AX88U stock firmware. Iâm finding that the integration is corrupting my static IPs and the router is âforgettingâ the name of my tracked devices and their statix ip management gets messed up. Only happens to devices tracked through the integration.
Anyone else?
Hello,
AsusRouter cannot directly access the DHCP server table and cannot change the IPs set to static. Please, check your deviceâs DHCP settings (available at https://YOURROUTERIP:8443/Advanced_DHCP_Content.asp). Are they changed from what it was?
As per the âforgetting the name of tracked devicesâ, I would need more context. How do you see this? When and how it happens? And what do you actually mean by the device name?
Thanks in advance for the provided info
What are your thoughts on adding a service to reload the integration? I see some integrations have this.
Just thinking that maybe reloading the integration on a daily basis may help with the device tracker issue. I know we donât want to rely on it as a permanent solution.
Isnât there a Service available in Core already that letâs you reload any integration? Why not just create an automation using that to do what you needâŚ
Is there? I looked through them but didnât see one that allowed you to pick an integration. Will look again. But yes, an automation using a service is what I am looking to do.
Something like this should workâŚ(does for me anyway, tho not with this integration)
Obv add in your own entity_id and/or device_id
service: homeassistant.reload_config_entry
target:
entity_id:
- entity1
- entity2
device_id: device 1
data: {}
Thanks! Iâll give that a try. I know itâs purely a workaround but interested to see if this helps.
Stumbled upon this integration after getting very frustrated with ASUSWRT. From what I can tell itâs FANTASTIC! I do have one question that Iâm close to pulling my hair out over! I have an ASUS AiMesh setup in my house consisting of:
1x Main Router
2x Wired Nodes
1x Wireless Node
I mainly use the integration as a device tracker so I can trigger certain events when family members are home vs. awayâŚHowever, my issue is that when a user (android or IOS) connects to the Wireless node (happens to live outside on the patio to cover the backyard), the entity state for that device shows as: AWAY even though they are still connected to the network. How do I tell this integration to over-ride also use the wireless node as part of the tracker?!
Thanks again for all the great community support!! Look forward to figuring out this small issue!
EDIT: I am running stock firmware and tried both Router and AP Access mode with no luck on either.
Hi Mark,
I also have an ASUS AiMesh AC set up in my house consisting of 3 units:
1 router
1 node (wired)
1 node (wireless)
My wireless node is also to provide coverage for the back yard however in still inside the house against the back wall. My wireless node only has the 5GHz radio turned on (2.4GHz disabled) as the backyard coverage I need is for family and guest cell phones. All my IoT devices (2.4GHz) connect to one of the two wired nodes either Ethernet of WiFi.
I too use this excellent AsusRouter integration to track which family members are visiting (adult children living away) as well as home residents (my wife & I) home or away via cell phones (combination of iPhones and Android devices). For me and my wife I also use geofencing apps to supplement the AsusRouter wifi presence.
I just tested the AsusRouter presence for myself by going into the backyard and confirming my iPhone was connected to the wireless node. My iPhone entity remained as âHomeâ and did not switch to âAwayâ when my phone WiFi connection roamed from the main router node to the wireless backyard node. I too am running stock firmware and the latest software release for the ZenWiFi AC hardware. Just thought Iâd share my findings and experience in case it helps you and/or the integration developer.
The device tracker finicky-ness is a known issue but tough nut to crack. When you are seeing this behavior, if you reload the integration, does it immediately update to HOME?