I have Home Assistant as well as various devices on my home network that are queried frequently, as a result, it skews the Admin Console reporting which shows both local and internet traffic in it graphs
Expected Behaviour:
I would like to be able to exclude local network traffic from these graphs so that it gives a clearer picture of actual Internet queries and the percentage of ads being blocked.
Actual Behaviour:
At times, actual Internet traffic is overwhelmed by local traffic as shown in these pics:
Update:
So I found the setting to exclude clients being shown in 'Top Clients' on the Dashboard. And that does cut down the 'Clients (over time)' graph. But the 'Queries over the last 24 hours' and 'Total Queries' are still counting queries from that excluded client. These are just local DNS Server queries and I'd really like to exclude them, but can't find a way to do that.
Thanks for any/all help:
P.S: I also added *.asgard to the "Top Domains/Top Advertisers" box, but they are still being counted in the Dashboard's 'Total queries' counter and show up at the top of the QueryLog.
Here's an example of what my dashboard looks like after just a few days with being unable to exclude smart home local traffic. It's basically meaningless now:
Your dashboard is accurately reflecting the DNS queries received and processed. It may not be showing you the external traffic only, but it is not meaningless.
I submit that its meaningless because Internet Ads do not originate from the local network, therefore traffic between those nodes should not be included: Pi-hole 'A black hole for Internet advertisements'
At the very least, its misleading: Hmmm, only 0.5% of all my traffic is ads, looks like I really don't need Pi-Hole after all. I don't get enough Internet Ads on my network to warrant using this App.
That's not quite what the display is telling you. It is telling you that of the total 211,476 queries from the 14 clients, 1,106 of them were blocked by Pi-Hole. This does not infer they are all ads. Blocklists contain domains for advertising, telemetrics, tracking, malware, etc.
If you aren't seeing the ads you are trying to block, then the Pi-Hole is working as expected. That is the measure of "goodness", not the numbers shown on the dashboard.
You could have a single client repeatedly request a blocked domain, and this would far outnumber the remaining requests. This also would skew the overall statistics, but would not impact the ability of Pi-Hole to accurately block DNS requests, using your blocklists.
My advice - don't be too concerned with the overall stats on the dashboard. The real issue is - are the domains that I want to be blocked in fact blocked? Am I seeing advertisements? If you are satisified with the answers to those questions, then ignore the dashboard and be satisfied that your Pi-hole is working as expected.
Here are four screen shots of dashboards from four Pi-Holes currently running in my house, serving different clients on the network. From looking at these, I infer nothing about the "success" of each Pi-Hole. I typically only look at the total number of queries and clients, as a check that there is traffic to the Pi-hole. The block rates are all over the place, depending on the connected clients, and I don't pay much attention to them for this reason.
