No worries, the other poster was just wasn’t being helpful. And/or doesn’t understand statistics & databases, but I don’t care to speculate on that or to waste more of my time on them.

The setting above maxes out at 24h in stock builds, but can be extended beyond that if you are willing to recompile the FTL database with different parameters to allow for a deeper look back window for your query log. Even at that point, a second database setting farther down that page sets the max age of all query logs to 1y, so at best you’d get a running tally of up to a year. This would probably at the expense of performance for dashboard page loads since the number is probably computed at page load. The live DB call is intended for relatively short windows vs database lifetime.

If you want an all-time count, you’ll have to track it off box because FTL doesn’t provide an all-time metric, or deep enough data persistence. I was just offering up a methodology that could be an interesting and beneficial project for others with similar needs.

Hey, this was fun. See you around.

You know what? I’m gonna disengage here. You’re not hearing what I am saying.

#### MAXLOGAGE=24.0
Up to how many hours of queries should be imported from the database and logs? Values greater than the hard-coded maximum of 24h need a locally compiled `FTL` with a changed compile-time value.

I assume this is the setting you are suggesting can extend the query count period. It still will only give you the last N hours’ worth of queries, which is not what OP asked. I gather OP wants to see the cumulative total of blocked queries over all time, and I doubt the FTL database tracks the data in a usable way to arrive at that number.

Ah, well if you know differently then please do share with the rest of us? I think the phrasing in my post makes it pretty clear I was open to being corrected.

So, like a running sum? No, I don’t think so, not in Pi-hole at least.

Pi-hole does have an API you could scrape, though. A Prometheus stack could track it and present a dashboard that shows the summation you want. There are other stats you could pull as well. This is a quick sample of what my home assistant integration sees

That counter, I believe, for the last 24 hours. It will fluctuate up and down across your active daily periods

Unless I misunderstand your question, draw.io can be downloaded as a standalone Linux application and run locally.

Likewise, the Xfig package should he available in most Linux repos. It’s old, but good enough for a quick sketch.

edit: aha. My mistake. My eyes slid over ‘open source’ in the title*, and even still I hadn’t realized it was an Apache license.

* Whaaat, it was pre-coffee? Let the purest among us cast the first stone.

Sure that’s easy. All you need is an HX711 sensor load cell amplifier taped to the floor of the chamber where it sits under an edge of your bottle.

Full, 5 gal (US) jug or water weights 40lbs nominal. If you have the same bottom feed dispenser I have, it starts drawing air with about 16-20oz of water left in the bottom.

You’d want to calibrate first for the sensor without weight (tare), and then for a full bottle’s weight (40lb) accounting for cosine error of the bottle being slightly tilted by the thickness of the sensor. Or, tape a couple of shims at 120° offset from the sensor to level out the bottle.

Finally, note the weight of an “empty” bottle with that 18ish oz of water in it, which will probably be close to the zero value.

Now you can calculate % full by looking at the ratio of the current weight between your full/empty value as:

%Water = [ (W_full - (W-18oz -W_tare) ) - W_now ] / (W_full - (W_18oz - W_tare) ) x 100

I can’t speak to the quality of actual locksets or recommended any products, but LockPickingLawyer on YouTube has a number of short videos discussing smart locks. My impression is most of them are trash at just being a lock. The ones made by traditional lockset manufacturers probably aren’t generally good at being smart, but I’ve been wrong about major brands’ commitment to open standards before.

Assuming you can find a decent lock that talks Z-Wave, I think you’re on the right track with Zwave2MQTT and a USB dongle. I’d be squeamish about using a Pi specifically in a mission critical security control system for a couple of reasons (reliability, complexity, WiFi interference), but as long as you have keyed backup, it’ll probably be OK.

I’ll leave it to others to recommend the locks, but as I mentioned in another post her, most battery operated Z-Wave devices, in my experience, report their battery life. Most of mine seem to go from 100% to 70% to dead in about a day though, so accuracy might be hit & miss.

Maybe just leave one door with an old school keyed lockset as a plan B.

My battery operated Z-Wave sensors all report their battery life. You’d need to dig into the integration docs to know for sure but frankly I’d be surprised if they didn’t report it as an entity.

Unbound will take updates via API. You could either write exit hooks on your clients, or use the “on commit” event on isc-dhcp-server to construct parameters and execute a script when a new lease is handed out.

I used to selfhost more, but honestly it started to feel like a job, and it was getting exhausting (maybe also irritating) to keep up with patches & updates across all of my services. I made decisions about risks to compromise and data loss from breaches and system failures. In the end, In decided my time was more valuable so now I pay someone to incur those risks for me.

For my outward facing stuff, I used to selfhost my own DNS domains, email + IMAP, web services, and an XMPP service for friends and family. Most of that I’ve moved off to paid private hosting. Now I maintain my DNS through Porkbun, email through MXroute, and we use Signal instead of XMPP. I still host and manage my own websites but am considering moving to a ghost.org account, or perhaps just host my blogs on a droplet at DO. My needs are modest and it’s all just personal stuff. I learned what I wanted, and I’m content to be someone else’s customer now.

At home, I still maintain my custom router/firewall services, Unifi wireless controller, Pihole + unbound recursive resolver, Wireguard, Jellyfin, homeassistant, Frigate NVR, and a couple of ADS-B feeders. Since it’s all on my home LAN and for my and my wife’s personal use, I can afford to let things be down a day or two til I get around to fixing it.

Still need to do better on my backup strategies, but it’s getting there.

I use Porkbuns API. It’s not sophisticated, but it works.

Gandi changed their TOS and price structure last year, so I ported everything over to Porkbun for a small savings, but mostly as a big middle finger to Gandi.

If you’re gonna get banged that kind of cash for functions you’re already using, you may as well look at better registrars, and get better value for your spend.

Shop around.

I’ve been happy with the GE Enbrighten series, though they do require a neutral wire. I use the Z-Wave models, but I can’t imagine the ZigBee models would work any differently.

Neat feature of the Enbrighten line is they recognize several gestures including short tap, long press, double tap, and I think even triple tap. I configured mine so that short taps on/off toggle the lights between off and 100% on, long press on/off brightens/dims, and double tap on activates the preferred evening lighting scene.

I think Minoston are also a GE product line? Again, Z-Wave, but rated for higher wattage in my case. They also require a neutral wire.

I had a pair of no-neutral ZigBee switches sold as ‘Martin Jerry’ brand, but they were super dodgy. Terrible feel to them, and they made a sort of electrical whine when the fixture was on. I pulled them right out and sent them back.

I don’t know this for sure, but if you installed the mobile companion app thru the Store for your platform, then the integration is probably using its push API to reach your device remotely.

I’m >90% sure notification workflows find me even when I am remote and untethered from home when I notify the app on my mobile devices.

You could source a pair of gigabit media converters and a length of fiber on Amazon for about $100. Just use the media converters to extend the Ethernet port from where the Internet hands off in your house over to your office. You can affix the fiber along baseboards and up over door frames with adhesive cleats and zip ties, or those nylon staples on a nail they use to tack down coax cable.

If you’re willing to spend a little more on the fiber for a custom color, you can probably even order the fiber in a more neutral color than SMF yellow to blend into the trim better.

Seems It was actually 2024.6.1 that introduced the change. Not sure where in this release notes it said it was a breaking change, but

I’m all turned around. I updated the main post with the link to the PR and I had the right version from the get-go.

Oops! Good eyes, and thanks! Typing yaml on mobile is squinty business. Fixed in the main post

If you just want each physical interface on your server to participate in a single VLAN, set the corresponding switch port as an access port in the desired VLAN, and then configure each server interface as a normal untagged interface.

You would only do tagged frames (802.1q trunking) if you wanted to support several VLANs on the switch port.