#Is there a way to ping Nabucasa URL to check Connectitivity without remote ui sensor

1 messages · Page 1 of 1 (latest)

ebon nimbus
#

I've observed that the Remote UI Sensor is not always accurate, almost never is, sometimes the URL is not reachable and the HA sensor detects as "Connected". Is there a way to have the ping integration ping itself but via internet? I do not want to use a service like uptime robot, trying to get something self contained...

orchid imp
#

I just read your ticket I think

ebon nimbus
orchid imp
#

The url is pointing to your HA server, so this is what is not connectable or not offering the front end. As that connection is initiated and maintained by HA itself, it is possible to have it stuck with showing a connection and the front end isn’t loading. Does it appear to hang and then time out when you use the url?

#

And yes, the Alexa and Google integrations are a separate connection

ebon nimbus
#

At the time that i pointed in the ticket, it says the instance is offline

#

doesn't offer the login page neither start loading

#

just shows the retry button

#

After this two hour period, everything was fine, it's an intermittent issue only

orchid imp
#

It would be good to see debug logs and host logs when this happens. Usually a full reboot is better than a restart of HA for this. The network adapter and network services are fully restarted with a reboot vs a restart of HA

ebon nimbus
orchid imp
#

There aren’t really intermittent issues on the Nabu Casa end

ebon nimbus
#

but if I'm far away from the house, the remote ui sensor doesnt update

#

I can't run an automation

#

Let me check the logs real quick to see if I see anything pop up

#

I have more than 10 nabucasa instances, and all of them occurs, as I said, nothing absurd, it's just strange and intermittent

#

actually very rare

#

But i've seen on all of them

#

it's different ISP, different houses, cities

orchid imp
#

Do they have hardware in common? Network hardware? ISP?

ebon nimbus
#

can't see any relation other then being on the same country hahahaha

orchid imp
#

Sorry I was too slow and your were too quick

ebon nimbus
#

Most of them are intel NUC, i think 3 or 4 are the Beelink PC

#

all of them using haos

orchid imp
#

I do see this issue but have never experienced it myself, and I’ve been using Nabu since it was started. Different installations over the years: RPi, Odroid, VM, Yellow…different network hardware has been Unifi and Omada. All HAOS.

ebon nimbus
#

2024-09-17 04:08:30.228 ERROR (MainThread) [snitun.multiplexer.core] Ping fails, no response from peer

#

Only log at the issue time

orchid imp
#

Host logs?

ebon nimbus
#

yes

#

next log is 6AM about an automation, unrelated

#

and the earlier 2AM about an automation too

orchid imp
#

If you check the system details on the repairs page using the three dots in the upper right corner, what do you see for your remote server and relayer region?

ebon nimbus
ebon nimbus
ebon nimbus
orchid imp
#

Same as mine. Well I have us-east-1-3

ebon nimbus
#

I’m from brazil, I don’t know if thats relevant

orchid imp
#

Are the routers the same for all of these?

ebon nimbus
#

i've seen 1-5, 1-3, 1-7

#

but relayer is the same

#

1-12 too

orchid imp
#

Any custom integrations in common or themes?

#

Are you managing these for other people (if you don’t mind me asking)

ebon nimbus
#

in common, i'll say, alexa media player

#

zigbee2mqtt

ebon nimbus
ebon nimbus
#

I've been testing the remote home assistant integration, but not all instances have them

orchid imp
#

The iOS theme has been causing issues lately

#

They issues really increased in number after the last update

ebon nimbus
#

Actually, i have a branch of them

#

I only use the colors as a base

#

to use my own theme

orchid imp
#

You might try using the default theme on one of these instances an see if it behaves better

ebon nimbus
#

I thin one of my friend's have the default theme still, but still we saw the issue

#

let me confirm

#

About the ping, is there a way I can ping my instance?

#

other than trying to open the browser

orchid imp
ebon nimbus
#

I mean

#

I want to auto check via HA

#

so that when this issue occurs, the host reboot by itself

eternal wolf
orchid imp
#

This concept where the intention is to have the HA server ping itself to see if it is connected has a flaw...

eternal wolf
# orchid imp This concept where the intention is to have the HA server ping itself to see if ...

I presume the HA server would not ping itself, but would ping something that's reachable via the network whose connection needs checking. The downside to that is that it will also flag a failure if it's the "something else" that's down, rather than the connection from HA to the network.

As I understand it, the intention is to determine specifically whether there's a path between the HA server and nabu casa's cloud server. The logical answer to that would be to ping the nabu casa cloud server. But that (in my case at least) doesn't respond to pings 😦

@ebon nimbus Probably the best you could do would be to manually do a tracert to your cloud server, and note the IP address of the node nearest the destination that does respond to pings. Then periodically ping that IP address to determine whether you have a connection to the network.

It's not ideal, because the path to the cloud server could change, in which case you would get into an infinite reboot loop so you would probably need to limit the number of reboots and flag an alert if that limit was exceeded. But it's the best I can think of as an improvement on doing no tests.

orchid imp
#

There is no instance of the HA server hosted in the cloud to ping.

eternal wolf
#

But there are plenty of other sites that can be pinged by the HA server if the intention is to determine whether the local HA server has a working internet connection.

violet summit
#

maybe a dns server could work too, 1.1.1.1 or something

edit: ah that has a website too so itll work

ebon nimbus
#

sometimes, using nabucasa

#

HA has internet connection, pinging google dns or cloudflare for example

#

but Remote UI still not accessible, making things worse, the remote ui sensor report as connected.

orchid imp
#

If it's in this kind of state, there's no proper communication happening. And if it is happening often, you might check the ha logs and the host logs