I have been having this oddly annoying problem: Client 192.168.1.20 has been rate-limited (current config allows up to 1000 queries in 60 seconds)
It affects either my smartTV, my WFH laptop which I do have a lot of network activities (IT), and when that happens the whole network goes to shit like literally Shit with capitol S
By reading other posts, this seems to be a new feature, I never added RATE_LIMIT=1000/60 into my setup and yet it appeared there.
In my experience this added no benefits but headache only and the log provides no useful information.
Instead of a raw has been rate-limited (current config allows up to 1000 queries in 60 seconds it should provide what calls were made like X domains to help you to identify if that was an issue indeed or nothing to worry about.
I am disabling this, I wished I looked into this sooner coz the whole network goddamn, but I would like to know, what is the goal of this feature?? It cannot be to add problems.
here's a good earlier thread about rate limiting. It has information on why rate limiting is important to note and how to change it.
In this thread, in 2021, @yubiuser wrote this-
In certain configurations it might be reasonable that your router has such a high number of queries, esp. if all your client traffic goes to your router first and then to Pi-hole. But you should have a look at your queries first to rule out a DNS loop or a client going wild. Do you always see the same queries in the query log?
You can increase the limit via /etc/pihole/pihole-FTL.conf. See the link mibere posted above.
The solution is not clear for me, I mean, the TV is very likely spamming Netflix and Samsung servers so that my explains why it is being limited.
But the documentation is clear, it affects that device only but in reality it affects the whole network. I have to restart both PiHole instances to have the network back without pages taking an eternity to load.
I am making a few API calls and having my laptop limited is a big no.
For the moment I have that disable like it used to be and hopefully no more network issues.
Yes. You can disable it both in the conf file and also in the gui. Just change it to 0/0. But check the logs over the next few days to make sure everything is ok.
I'm nowhere near a developer but in guessing they included this originally to not overload the Pi and also to help people see that there may be a situation that needs to be attended to.
What is the device at 192.168.1.20 which is being rate-limited? The Query Log will let you see what this device is doing.
If you want, please create a debug log and post the token URL here and someone can take a look at your environment. You can use the command pihole -d or use Tools > Generate debug log and select the relevant options.
The rate-limiting of one device should not be affecting the performance elsewhere on your network. The exceptions to this might be if the Pi is a low spec early model and is being heavily loaded, or if the rate-limited device is your router and is therefore impacting your network. There are ways around both of those scenarios.
That is my WFH laptop which I have been using to call a bunch of APIs but not 1000 calls within a minute.
The information provided under "Pi-hole diagnosis" which is how you find out a deice is rate limited, it doesn't help much.
I mean, it tells you which device is but it lacks a proper report like "Okay, your laptop called this domain hundreds time..." or I am missing something.
I don't run RPi but Debian VMs on Proxmox, they are like 1 core 2GB. Under normal loads the CPU is used less than 2% and memory barely reaches 200MB used so environment isn't a problem.
I have two instances, primary and secondary DNS, they both behaves exactly the same.
No need for the ticket support, I have just disabled it altogether. Will try to use NTOPNG to understand what the devices are calling so much. So far the network performance is as good as I know it for.