32:39NetworkChuck · Tutorial
I'm switching to Hermes (goodbye OpenClaw!!)
A 32-minute live walkthrough where NetworkChuck installs a self-improving AI agent, names it Ron Weasley, and never looks back at OpenClaw.
May 20thThe argument in one line.
One free monitoring server can watch every machine, container, VM, website, and script you own, because Checkmk collects data six different ways and any of them can be bent to monitor something it was never designed for.
Read if. Skip if.
READ IF YOU ARE…
- You run a home lab or a small-business rack and find out things are broken because a person tells you, not because software did.
- You already self-host Docker, Proxmox, or a NAS and want one dashboard instead of five separate web UIs.
- You have been quoted per-host or per-service pricing by a commercial monitoring vendor and want to know what the free tier actually covers.
- You run AI agents or long-lived scripts on their own machines and have no idea when one silently stops responding.
- You want to understand how agents, SNMP, API integrations, and piggyback data differ before you pick a monitoring approach.
SKIP IF…
- You want a zero-config hosted dashboard. The free edition here needs manual plugin installs, rule-writing, and CSV imports to reach full coverage.
- You only have one machine. The setup overhead only pays off once you have several things that can break independently.
- You need log aggregation or distributed tracing. This is host and service state monitoring, not an observability platform for application traces.
The full version, fast.
Checkmk Community edition installs on Ubuntu with two commands and gives unlimited hosts and services for free. The core loop never changes: add a host, pick how Checkmk will reach it, run service discovery, accept the undecided services, activate pending changes. What varies is the collection method. An installed agent gets deep local data, SNMP sweeps a subnet for network gear, a special agent talks to an API like Proxmox without installing anything, piggyback data lets one host report on children such as Docker containers, and active checks run scripts from the monitoring server itself. Anything that prints a status line can become a monitored service, which is how custom Plex, Home Assistant, and AI-agent checks get built. Alerts then route to Slack or to an automated ticketing agent.
Free for members
Chat with this breakdown — free.
Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.
Create a free account →Where the time goes.
00:00 – 00:53
01 · Cold open
A rapid list of everything that could be down right now, followed by the promise: one free tool that monitors all of it, routes alerts to an AI IT department, and ends up configured by an AI agent.
00:53 – 03:00
02 · Install in five minutes
Download the Community edition package from the Checkmk site, install it, verify with omd version, create a monitoring site, start it, and log in with the credentials the installer printed.
03:00 – 07:16
03 · First Linux host and the agent
Add a host by IP, install the packaged agent on the target, open the firewall port, test the agent connection, run service discovery, accept 44 undecided services, and activate the changes. Discovery immediately flags a dead NFS service.
07:16 – 10:49
04 · Network gear via SNMP
Instead of adding devices one by one, create a folder with a network scan over an IP range and SNMP credentials. Checkmk sweeps the subnet, finds the MikroTik switch and UniFi gear, and discovers interfaces, CPU, and uptime on each.
10:49 – 16:47
05 · Docker containers via piggyback
The longest and fiddliest section. Install the agent on the Docker host, manually add the mk_docker.py plugin plus its config, then create a second folder set to no IP, no agent, no API, always expect piggyback data, and bulk-import the container names from a CSV so each container becomes its own host.
16:47 – 20:24
06 · Proxmox through the API
Monitor the hypervisor with no software installed on it. Add the host with API integrations only, write a Proxmox VE rule with a dedicated username, password, port 8006, and SSL validation disabled, then scope it to that host. Renaming the Docker VM to match its Proxmox name links the two and pulls in snapshot age and VM memory.
20:24 – 23:55
07 · Websites and a Windows PC
An HTTP web service rule checks a public site for status codes and an expected string in the body, proving the alert by searching for a word that is not there. Then the Windows agent goes on the recording machine, which immediately reports a near-full file system, a 10 gig NIC running at 5 gig, and a stopped time service.
23:55 – 27:18
08 · Plugins: Ceph, drives, and the Exchange
Adding the Ceph and SMART plugins exposes the four-node NAS cluster and per-drive temperature, revealing an SSD pool at 97% during the recording. Then a UniFi controller plugin is pulled from exchange.checkmk.com, installed as an .mkp package, and enabled by name.
27:18 – 30:08
09 · Custom checks for anything
An AI assistant writes a Plex plugin from a spec, wired in through a Nagios plugin rule as a command line. Home Assistant gets the same treatment with camera, entity, light, and motion counts. Then the AI agent machines themselves are monitored through local scripts that the installed agent executes and reports as services.
30:08 – 32:23
10 · Alerts to Slack and to an AI ticket queue
A notification rule sends state changes to a Slack webhook, demonstrated live by stopping an agent gateway and watching the alert and recovery arrive. A second rule filtered to one service calls a script that opens a ticket with an automated IT agent, which restarts the gateway on its own.
32:23 – 34:40
11 · Handing the keys to an AI agent
A user is created with an automation secret and administrator role, giving machine access to the REST API. The AI agent then adds every remaining host and builds a custom dashboard, and the final tour shows host and service problem views, Docker containers, and a per-agent dashboard with two critical findings.
34:40 – 37:53
12 · Final thoughts
A note that the Community edition's free coverage is unusual, that the Ultimate trial stays free up to 750 services on one site, the usual subscribe ask, and a closing prayer that reframes monitoring around the parts of life nobody instruments.
Lines worth screenshotting.
- Checkmk Community edition has no host or service limit, which is the opposite of how almost every monitoring vendor prices the free tier.
- The entire workflow is one repeating loop: add host, choose collection method, run service discovery, accept undecided services, activate pending changes.
- Checkmk stages config changes like code and makes you activate them, so you can review a batch of edits before anything goes live.
- An installed agent and an API special agent can both run against the same host, because each sees the machine from a different angle.
- Piggyback is the trick for monitoring things that have no IP of their own: the parent host reports on its children and Checkmk files the data under separate host objects.
- Monitoring Docker containers in the free edition means hand-installing the mk_docker.py plugin, creating a no-IP piggyback folder, and importing container names via CSV.
- SNMP is already running on most routers and switches, so a folder with a network scan and a community string discovers your network gear without touching any device.
- A Proxmox host is monitored entirely through its REST API on port 8006, with no agent installed on the hypervisor at all.
- A VM only links to its Proxmox parent when the Checkmk host name exactly matches the VM name on the hypervisor.
- An HTTP check can search the response body for an expected string, so you get alerted when your site is up but serving the wrong content.
- Any script dropped in the agent's local checks directory becomes a monitored service, which means anything you can program, you can monitor.
- Active checks run from the monitoring server outward, so you can watch a system that will never let you install software on it.
- The Checkmk Exchange is a community plugin marketplace, and plugins install as .mkp packages that you add and then enable by name, not by file name.
- Checkmk collects every metric once a minute and retains it for up to four years, with default good and bad thresholds already baked in.
- Turning monitoring on surfaces problems immediately: a 97% full SSD pool, a 10 gig NIC negotiated at 5 gig, a stopped Windows time service, and a switch port up for 405 days.
- Checkmk notification rules can call an arbitrary script, so an alert can open a ticket with an AI agent instead of pinging a human.
- Creating an automation secret user gives an AI agent full REST API access to add hosts, write rules, and build dashboards on its own.
- The Ultimate commercial trial stays free for up to 750 services on one site, which is more than most home labs will ever need.
One server, six ways to see a machine
WHAT TO LEARN
Monitoring everything is not one feature, it is five collection methods applied to whatever each target will tolerate, plus a repeating six-step loop you run until nothing is invisible.
02Install in five minutes
- Put the monitoring server somewhere on the network that can reach every subnet and device you intend to watch, because collection is pull-based.
- The Community edition installs from two package commands, and creating and starting a site is one command each, so the install is not the hard part.
- The web interface lives at the server's IP followed by the site name, and the installer prints the default admin credentials to the terminal.
03First Linux host and the agent
- An installed agent is the deepest source of data on a machine, reporting file systems, services, interfaces, and process state that ping can never reach.
- Service discovery lists candidates as undecided services, and nothing is collected until you accept them and activate the pending change set.
- Config edits stage like code and need an explicit activation, which means you can batch several risky changes and review them together before they go live.
- Expect roughly five minutes before data appears, because collection runs on a one-minute cycle and graphs need several points to be readable.
04Network gear via SNMP
- A folder with a network scan plus SNMP credentials discovers a whole subnet at once, which is faster than adding switches and access points by hand.
- SNMP already runs on most network hardware but is usually disabled, and the community string on the device must match the credential in the rule.
- Discovered hosts arrive flagged as do-not-monitor, so you have to bulk-edit the attributes before anything is actually watched.
05Docker containers via piggyback
- Docker containers need the mk_docker.py plugin installed by hand on the host, along with python3-docker and a config directory the agent knows to read.
- Containers have no address, so they get a folder set to no IP, no agent, no API, and always expect piggyback data from their parent host.
- Container host objects are created by pasting the container name list into a CSV import with a host_name header, which is the step most people miss.
06Proxmox through the API
- A Proxmox host is monitored through its API on port 8006 with a dedicated read account, so nothing is installed on the hypervisor itself.
- Running the installed agent and the API special agent on the same host is encouraged, because each one sees the machine from a different angle.
- A VM only inherits its Proxmox context when the host name in Checkmk matches the VM name on the hypervisor exactly, including case.
07Websites and a Windows PC
- An HTTP check can assert an expected string in the response body, so you catch a site that returns 200 while serving defaced or wrong content.
- The Windows agent installs as a normal package, but a blocked test connection usually needs both a firewall rule and a restart of the agent service.
- Turning monitoring on a machine you thought was fine surfaces real faults immediately, such as a link negotiated at half speed or a stopped time service.
08Plugins: Ceph, drives, and the Exchange
- Plugins are what expose specialist subsystems, and without the Ceph and SMART plugins a storage cluster looks like four ordinary Linux boxes.
- The Checkmk Exchange hosts community plugins as .mkp packages, added with one command and then enabled by package name and version rather than file name.
- Marketplace commands have to run as the site user, so switching user before adding a package avoids the command-not-found dead end.
09Custom checks for anything
- A custom check is just a script wired in through a Nagios plugin rule, which makes any API with a token monitorable in an afternoon.
- Scripts dropped in the agent's local checks directory are executed automatically and appear as services, which is how long-running agents get heartbeats.
- Active checks run from the monitoring server outward, so you can watch systems that will never let you install anything on them.
10Alerts to Slack and to an AI ticket queue
- Notification rules filter by host and service before choosing a delivery method, so you can route one specific service somewhere different from everything else.
- Testing an alert by deliberately stopping a service proves both the alert and the recovery path, which a test button alone does not.
- A notification method can be an arbitrary script, so an alert can open a ticket in an automated system instead of waking a person.
11Handing the keys to an AI agent
- A user created with an automation secret and an administrator role gives a script or agent full REST API control over hosts, rules, and dashboards.
- Delegating bulk host-adding to automation only works after you have done it manually enough times to recognise a wrong result.
Terms worth knowing.
- Checkmk agent
- A small program installed on the machine you want to watch. It runs locally, gathers detailed system data such as CPU, disks, file systems, and services, and hands it back to the monitoring server when asked.
- Service discovery
- An automatic scan that asks a host what it can report on and returns a list of candidate services. You then choose which of those candidates become monitored services.
- Undecided services
- Services that discovery found but that nobody has accepted yet. They exist in the config but generate no data or alerts until you promote them to monitored services.
- Activate pending changes
- A commit step. Edits to hosts, folders, and rules are staged in a change list and only take effect once you review and activate them, similar to committing code.
- SNMP
- Simple Network Management Protocol, a long-standing standard that routers, switches, printers, and access points use to expose their status. It usually needs to be switched on and given a community string, which acts as a shared password.
- Special agent
- A connector that runs on the monitoring server and queries a system's own API over the network, so no software has to be installed on the monitored machine. Used here for Proxmox and the UniFi controller.
- Piggyback
- A mode where one monitored host reports data on behalf of other things that have no network presence of their own, such as containers. Checkmk splits that data into separate host objects so each container gets its own page.
- Active check
- A test executed by the monitoring server itself rather than by software on the target, such as fetching a URL or running a script that queries an external service.
- Local check
- A custom script placed in a specific directory on a monitored machine. The installed agent runs it automatically and whatever status line it prints becomes a service in the dashboard.
- MKP package
- The Checkmk extension file format. Packages downloaded from the Checkmk Exchange are added with one command and then enabled by package name and version, not by file name.
- Automation secret
- A machine account credential type. A user created with an automation secret can authenticate to the Checkmk REST API without a browser login, which is how scripts and AI agents drive the configuration.
- Ceph
- Clustered storage software that pools the drives of several machines into one shared storage system. A Ceph cluster is often what sits behind a self-built NAS.
- Perfometer
- The inline bar shown next to each service row that renders its current value as a quick visual gauge, so a list of services can be scanned without opening each graph.
- Nagios plugin
- A script that follows an older but near-universal convention for monitoring checks: it prints a status line and exit code that most monitoring tools, including Checkmk, know how to read.
Things they pointed at.
08:00productMikroTik switch (SNMP source)
09:00productUniFi controller and switches
24:00toolCeph clustered storage
27:40productPlex Media Server
28:20productHome Assistant
30:30toolSlack incoming webhooks for alerting
Lines you could clip.
04:30
“He's like the inside man, just sitting there ready to tell us all the secrets of that server.”
14:00
“I told you, we can do everything, but it requires just a bit of work. We gotta dig for that gold.”
24:15
“Dude, look at my SSD pool. 97% full. I have to offload this video recording right now to this.”
30:00
“So whatever you can program or script right here, you can monitor anything, anything.”
10:50
“Hey look, there's Terry's port. It's been up a year and 108 days. Oh my gosh. I did not know that.”
32:10
“Ron just finished. Let's see what he did. He restarted the gateway.”
Word for word.
Read-along
Don't just watch it. Burn it in.
See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.
metaphoranalogystory
Hey, is your stuff down right now? Is your Linux server up? Is that Hermes Gateway responding?
How about that vibe -coded project you're super proud of? Websites? Is your NAS full?
Is that hard drive about to fail? Docker, Proxmox, Unify, Plex, Home Assistant? There is so much stuff to monitor now, but I just found a way to monitor...
everything with this free open source software. I can monitor every AI agent service VM container. I can literally monitor anything and everything in my home lab.
I can't believe this is free. So in this video, we're gonna set up check mark and have it monitor every single thing in my home lab, even my toilet. I'm not kidding about that.
Watch watch later. And we're not gonna stop there. We're gonna have my alerts sent to my AI IT department and they will solve my problems for me.
And then we'll get crazier. We're gonna connect my agents to the check mark API so they can configure and monitor things on demand. So get your stinking coffee ready.
It's time to monitor everything.
Now we're not gonna waste any time. We're gonna install check mark right now. You and me.
And it's gonna take us less than five minutes and a massive shout out to check Mark for sponsoring this video and just saying, Hey Chuck, show them our cool free tool. That's it. Help them monitor all their stuff.
So let's do that. Wait, actually it's pronounced check MK. I just learned that.
And actually real quick while check MK is amazing software and will monitor pretty much anything you want. It does take some tinkering with the free version to make all that possible. I'm gonna show you all that.
Now we're deploying this check mark monitoring server in our home lab. So keep in mind, you wanna put the server in a place where it can see all the stuff you want to watch. And here are the system requirements.
Boom. Thank you editors. And here are my specs.
Boom. I'm installing mine on a virtual machine. You can install it with Docker or via a package.
So here we are at my VM. Don't blink. It's gonna be fast.
Actually first, the first place we're gonna go is to check Mark's website. Then we'll go to downloads. I'm choosing Linux Ubuntu 2 ,404 is my current version and we're doing community edition.
We'll copy this command right here. Then make it super easy. Paste that into our terminal, download that little coffee break, and then we'll install the package with this command right here.
We'll just copy that paste that sucker go. Don't wanna continue. Yes I do.
And it's done and just make sure everything went perfectly. We'll run this command right here. We'll do OMD version.
That looks good to me. And then we're going to create a check mark monitoring site with this command right here. We'll copy that paste it here.
Pseudo OMD create monitoring. It's creating a monitoring site. And by the way, this is probably the hardest part of this entire tutorial and it's easy.
It's done. Like let's go visit our website, our new monitoring site. Here's my IP address.
I'm gonna copy that. And actually real quick before we do that, we need to actually start the server with the pseudo OMD start monitoring command and we're done. So now with it started, we'll go out to our IP address forward slash monitoring.
There's check mark. And then we'll log in with the credentials that told us where in the terminal right here, there's your default admin user account and your password. Don't steal mine.
And we're in, I told you it was gonna be fast. Now we're here. They do have a really fun tutorial on how to add all your stuff.
You don't need that. That's why you have me here. So just ignore that for a bit.
Now our first goal is to add a host that we can monitor for me. I'm gonna add a Linux server and it's actually a special one. You'll see why here in a moment.
So to add our first host, we'll go to set up over here on the left, click on that. And then we'll click on host right here.
And then finally add host right here in host name is where you'll add the either DNS name or IP address of the thing you're trying to monitor. For me, it's gonna be a host IP address of 10 .77 .14 .14. But hold on one second.
What does it mean to monitor this Linux server? Like how do we do that? When check mark, there are six ways we can do it.
Now you might be familiar with something like ping. For example, if I were to open up my terminal actually from our monitoring terminal and type in ping, 10 .77 .14 .14.
My Linux server. I'm now seeing if my Linux server is up. I'm saying, Hey, are you awake?
And he's responding. Yes, I am over and over and over. Now that's great, but I kind of wanna know more than that.
The best way to do that is with the check mark agent. An agent is something we install on the server, the thing we want to monitor, and it's going to help us figure things out about it. Here's how we do it.
Now I'm gonna open up another check mark window because I'm not done adding my host here. And then right down here, you'll see a section called agents, an agent for anything you could possibly want. And what we care about right now is Linux.
Click on that. A lot of stuff here, but all we care about is this section right here, the packaged agents. We have one for Debian based distros and one for RPM or red hat based distros.
Mine is Rocky Linux. So I'm gonna grab this. And what I mean by grab is I'm going to right click this link here and copy link address.
Then I'm gonna log into my server. We'll type in curl. dash capital O and then paste in our URL hit enter.
It's gonna download it. It was done. That was quick.
And then for me, I'll do DNF install dash Y period forward slash in the package. I just downloaded. If you're on something like Ubuntu, it might be APT install.
I'm going to install it. Ready, set, go complete. And then I'll add this firewall command to make sure I open up the port that my check mark server will need to access this server.
Now that was it right now, there is an agent living on this Linux server. He's like , just sitting there ready to tell us all the secrets of that server. So now the agent installed, we just have to tell check mark that it's there.
All we had done here so far is told it, Hey, this is the IP address. I want you to monitor, but also notice down here under monitoring agents, we have the option to test an agent connection. Let's try it.
We just added one. So it should work. Click on test fingers crossed.
Yes, we got it. And really that's all we need to do. Just go up here and click on save and run service discovery.
And this is actually a pretty fun part. Click on that. What this will do is auto discover everything about our server right now.
Check mark in that inside man. They're talking, look at all this stuff. Now what happened here is actually pretty sick checkmark through its auto discovery, trying to figure out what's on my host, use what's called plugins.
They're right over here on the right. And these plugins can expose the services they want to monitor. Like the heartbeat CRM resources plugin found that I have a critical NFS service down.
Dang it. But just look at all the stuff it found. These are all things that we can monitor on our server.
So I'm pretty happy with that. So notice right now, these are all undecided services currently not being monitored. I wanna monitor all these.
So at the top here, we can say, you know what? Let's do that. Monitor undecided services.
Boom. Cool. They're now our monitored services.
Notice down here on the left, we have changes with a notification icon. We have three. Let's click on that right here.
It gives us a summary of our changes and it wants us to actually commit these changes. Like kind of like code. You're gonna see this whenever you do anything, add a new host or whatever.
So what do you say? We activate our pending changes right up here. Let's click on that activate.
Now don't check it just yet. Give it about five minutes. But five minutes later, boom, here we are.
Let's look at all hosts. And there he is. Let's click on him.
There's all the services. You can see we have the perfometer over here. I kind of like that actually, perfometer.
And we're starting to get data. If I click on any one of these, like let's choose the CPU utilization. Not a lot of time.
because we just started collecting data, by the way, it collects data every minute, every one of these things, every minute, it'll keep it for up to four years. And every one of these metrics has some default thresholds baked in on what's good and what's bad. So you can be alerted when something goes wrong.
Now I'm embarrassed to say I haven't been monitoring my stuff like this. So I'm actually really excited to add all this. Okay.
We got our first Linux host. Let's add three more ready set. And clearly I have some work to do.
While these are my Ceph cluster nodes, you'll notice that when I go into what's monitored, we're seeing some stuff, but we're not seeing everything. Like I don't see anything about Ceph, which describes what's happening with my NAS. But first all we've added so far are just Linux hosts.
I wanna see other stuff like networking. Can it monitor my network? Yeah, let's make that happen right now.
We're gonna go back to setup and host. Now notice so far we've like added individual hosts, but this time we're going to add a folder, add a folder. We're gonna label this folder networking, and this is gonna unlock something cool for us.
It's gonna save us some time. Do you see what's happening right down here? We have the option for a network scan.
Let's set that up, enable it. Add a new IP range. I'll add mine in right now.
And then the way we're going to discover our host and actually access what they're doing is with this wonderful technology right here, SNMP. You might recall that SNMP is one of the ways we can get access to our host and see what they're doing. .
SNMP, by the way, stands for the Simple Network Management Protocol. And it's a service that is likely running on your routers and switches right now. Everything pretty much supports it.
It's a universal standard, but you have to enable it. Let's go make sure it's enabled on my MicroTik switch. I'll get logged in and I'll enable SNMP.
Let's see if the settings took. Yes, enabled. With my community name, network check.
And then get me back to check mark. Check MK. I will enable SNMP right here.
SNMP v2 or v3. And then I'll put credentials. That'll be the community name.
network Chuck, and that should be it. So now what we can do is click on save and notice we have our folder here. No host just yet.
Again, we have to commit our changes. We'll go to changes, activate pending changes, but what it will do when it does scan is it will scan for every host in that subnet and attempt to use the SNMP credentials to discover anything about it. It'll also be able to ping host and see if it finds things.
And we'll see that happen here momentarily. And actually look at that already. We have six changes.
And these are, Hey, we found some hosts. Can we monitor these hosts? If we actually go to the folder.
So we'll go to set up host and go to the folder. We see they do show up there. And before we make the changes, I actually wanna say, Hey, I wanna monitor these things.
Now let's do that. Microtick. Yes.
I'll select this. There's my unified controller and a unified switch. That's all I'm gonna do right now.
And with these selected, I can make some bulk changes. So I'll go to host edit attributes. And down here under custom attributes, which I think is kind of funny.
These are custom attributes and they're saying, do not monitor this host. No, I do want to monitor this host. This is a productive system, which again is also a very funny way to say that production.
So we'll save that. And notice how that scary little sign is gone. Now on these hosts.
Next, what I'm gonna do is discover what's happening with these devices. So for example, my 24 port unify switch, I'm gonna select that. And then at the top here, I'll click on save and run service discovery and click on rescan.
It's gonna use the SNP credentials and scan this host. And I do have SNP enabled on that switch. And I do have the community network chuck set up.
That's kind of the authentication, the community string. And it found look at that. everything interfaces, CPU utilization, easy for me to say uptime.
And it looks like it already made these monitored services for me. So we're good to go. I'll do the same thing for the other host as well.
Like for example, this unified controller has some undecided services. I'll accept all. Okay.
Now to accept all of our changes, go here. I will activate pending changes. So now let's go see if our network stuff showed up.
We'll go to monitor, go to all host. There they are. And like most things you have to give it a second.
I can see that things are being well seen and monitored, but you gotta give it a moment for it to find stuff. And that's, that's the fun thing. So we're gonna stop for a moment, take a little coffee break.
They have some stuff. Let's click on the MicroTik one. Look at all this stuff.
I love it. All the graphs on the side, the perfometer. Ooh, our journey just started here, but we can see everything.
Hey look, there's Terry's port. It's been up a year and 108 days. Oh my gosh.
I did not know that. I need to reboot that thing. We got Linux.
We got networking. Now what about Docker? Can we monitor our Docker host and actually see the containers?
Yes, we can. Now we'll start with my primary Docker host. And just like with our Linux host, we have to add an agent to it to be able to see all the goodies.
So I'll get logged into that Docker host. I will download the check MK agent this time using the dev agent for Debbie and based environments. And then I'll install it.
Sudo APT install that agent thing. And go cool. Now we can add our host to check mark, check MK.
I'm never gonna get this right. Just like before set up host, add a host. I call it Hogwarts Docker zero one.
I notice that name does not resolve. That's okay. I can actually add an IP address down here and I should make that shut up much better.
That's all I need. I'll click save and run service discovery. And it found everything except Docker.
It's not seeing anything Docker. Why? because we need a plugin.
You might recall that we have our plugins over here, our check plugins that will find services we can monitor. All these right here are default and will be discovered auto magically with the community version. What will not be done auto magically is the Docker plugin.
It is here and we can use it. We just have to do a bit of work. That's the price of free.
So real quick, I'll say monitor undecided services, except all, but then I'll go to setup. I'm gonna make this bigger for us and under agents, I'll select Linux. Yeah.
I'm gonna search. I'll do control F search for Docker. And there it is actually not that one.
This one, this right here, MK underscore Docker dot P Y is the Docker plugin built into check MK that we can add to our host, but we have to manually add. So we'll right click that copy the link address, get back to our Linux server, our Docker host. Actually first we do need Python three Docker installed.
So we'll do pseudo APT install Python three dash Docker.
Then we will download that plugin, this command curl dash O. I'll tell it where to go. And we're gonna put this in the place that Checkmk will look for, or rather the Checkmk agent will look for plugins.
That's where we're downloading it. And then here's where we're downloading it from. We'll add some pseudo action before it.
Download it. We'll then do a chmod to make it executable with the plus X. I'll then use this command, and this is where things get kind of complicated.
I told you, we can do everything, but it requires just a bit of work. . But this command right here is making a new directory for our config.
And we're pulling down this Docker example config from our host. See, there's the Docker config file as it continues over here. And we're downloading that Docker config file.
Ready, set, go. Failure, I need some sudo action probably. Sudo before make directory possibly, and sudo before curl.
There we go. And then I'll run this command to get a list of all my Docker containers that I have on the server. Keep that there.
We're gonna come back to it. It's important. You'll see why again work.
We'll get back to check MK, go back to set up, go back to host, get to our Docker host, click on save and run to service discovery re scan. And then some magic should happen. Now it's gonna automatically find those Docker host and that Docker config, the plugin.
There it is. Ah, we did it undecided services, Docker containers and Docker node info. We're not done yet.
We're gonna monitor these undecided services. Click that little button up there. Actually it's kind of a big button except all.
And while we're here, let's go ahead and go to changes and activate all these changes. Cause it's making me nervous. We haven't done that yet, but we're not quite done yet.
Cause we should be able to go to monitor and over here at applications, we should be able to see our Docker containers. But when we click on that right now, there's nothing let's fix that. So we'll go back to set up.
We'll go back to host and we're gonna create another folder right here. Add folder. And we'll call this Docker containers.
Click on show more. You're gonna see everything we wanna mess with right now for IP address family. We'll click on this and say no IP.
I know it's weird. And then over here with check MK agent API integrations, we'll check this and we'll say no API integrations and no check MK agent. So it's like, don't check anything.
Don't look anywhere. But then here's the secret sauce piggyback. we're gonna select this and we're gonna change the setting to always use and expect piggyback data.
Now, what the heck does that mean? It means that it's going to rely on another host to get its information. It's gonna rely on the Docker host that we've already set up.
It's gonna piggyback off that data to pull in these containers. Now we'll click on save at this point. Now let's get back into that Docker containers folder.
And then at the top here in our menu, we'll click on host. And we're gonna import host via CSV file. Click on that.
We'll make sure the import host option says content of CSV file. And then here is where we post the names of those Docker containers that we got from our terminal just earlier. I'm gonna grab these real quick.
Copy them, paste it right here. And after the first line, we need to add a header. And it's gonna be host underscore name.
I'll say perform automatic service discovery. And I'll click on upload. And we should see all of our stuff listed here to make sure it all looks good.
Click on import and then start progress info. If found all of them. Perfect.
And notice we have a bunch of changes coming in. Now, again, this is a free community version thing only that we had to do here. The paid version, it takes care of a lot of that for you.
Okay. Quick correction. Check MK actually has three commercial additions, not just one there's pro there's ultimate, which also comes in an MT version MT for multi -tenancy.
And then there's cloud, which is their SaaS platform. Go to our changes, activate our pending changes and we'll give it a second little coffee break right now. Now we should be able to go to monitor.
Let's look at all hosts and there those Docker containers are right now dashes down. But also if I go to monitor now and go to applications where you are, there you are. And then Docker containers, ah, they're right there and they're up and all that data is coming in.
Let's click on one of them. See what we can see. Uh, let's do engine X.
I use that one. I depend on that. We got CPU disc IO status of the container memory of time.
Like that's pretty fire. I like this. I'm excited.
We can actually see our containers. Can you contain yourself? You, you saw it coming.
You know, it was gonna happen. No, I do have another Docker host with a bunch of Docker containers, but I'm not gonna add that myself. I'm lazy.
I'm gonna have AI do that for me later. Can we actually monitor prox mox? Of course the prox mox host themselves, but what about the con not containers LXD containers and also the VMs.
Can we see the individual VMs? Let's make that happen. The Proxmox is actually special the way it's monitored.
And actually it's gonna use this, the special agent API. We haven't talked about that yet. It's kind of exciting.
It's very exciting. Now, first we're gonna add our Proxmox. We'll go to setup and host click on add host.
I'll name mine. Hogwarts Proxmox one. It won't resolve.
I'll add the IP address down here. It'll check. We're good.
And then under here, under monitoring agents for check MK, I probably said check market million times at this point, API integrations. We're gonna check that and say configured API integrations, no check MK agent. So we're officially not installing anything on Proxmox.
We're going to be integrating with this API and that's the special agent. Oh, and here's a pro tip. You don't have to pick one or the other run the check MK agent and the special agent together on the same host.
Each one sees that machine from a different angle. So you get the full picture. with this scenario, we're not putting a secret agent inside the system.
That's , telling us stuff. No, we're not doing that. Check MK is gonna reach out to the API and go boom.
Like you already have these doors open. I've got a username and password. Can you tell me about this?
Let's set that up right now. So actually first we need to save this and I'm not gonna click on save and run discovery cuz I don't wanna discover anything yet. I just wanna save and view folder.
So save and view folder. Cool. Hogwarts Proxmox is just sitting here waiting for us.
Now we have to set up that special agent. We're gonna go to setup once more and we're gonna go to the search box and we're gonna type in Proxmox. And we'll select this first option, Proxmox VE, this option right here.
Click on add rule, describe it however you want, Proxmox. And then under here, under Proxmox VE settings, click on username. And if you haven't already, you need to set up a Proxmox username and a Proxmox password for Checkmk.
So here we'll add in that username. inside of check MK, the password, the port, which will be by default 8 ,006. Then we'll say specify Proxmox VE host via IP address.
We're gonna set that setting. I'll disable SSL certificate validation because I'm doing self -signed certificates. And then under conditions, we're gonna specify which host this will apply to.
We're kind of setting a rule, like which of our hosts are actually Proxmox host. We'll leave everything the same except for clicking explicit host and then selecting our host name, which is Hogwarts Proxmox one. And we'll click on save.
That's our rule. Now we'll go ahead and go to our changes and commit all of our changes. So we're going to add that host, the Proxmox host we added, and also we're going to add this rule.
But now we're going to go back to setup, back to host, back to Hogwarts Proxmox one, and we'll do save and run service discovery again. And boom, it found all of our Proxmox info on this host. So yeah, we're going to monitor all undecided services, except all.
go over here and commit those changes or activate those pending changes. Now what's cool here. Watch this.
If I go to monitor and I'll host this host right here, my Docker host is actually a VM on that server, but I have to make sure the name matches what's actually on the server. And it's not matched right now. I need to edit this.
So I'll change the host name to Hogwarts Docker zero one all lowercase. Yes. Rename.
I'll go back to set up and host, get back into my Hogwarts Docker zero one and do a full rescan. And now it recognizes that it is indeed a Proxmox VM and it includes that data and memory usage and everything. It's so cool.
Snapshot age. I mean, come on, no snapshot found. We should fix that.
But anyways, monitor all undecided except all push these suckers. These changes get already. We're monitoring a million things.
Proxmox, Docker, Linux host. I mean, oh, it's awesome. And what about websites?
Websites? Can you monitor your stuff on the outside? Yes.
Let's make it happen. We'll go to setup. We'll go to host.
We've been here before. We'll add a host. I'll name this network chuck.
It can resolve that one. We'll see. We don't want any check mark agents or check MK agents or API.
None of that. And I'll just say save and view folder. Cause we're gonna do that weird thing where we add a rule.
So we'll go to step and search for HTTP. There it is. Check HTTP web service right here.
Click on that. We'll add a rule. And I'll just call this HTTP stuff, because this could possibly, you know, check multiple hosts.
And then under HTTP web service endpoints to monitor, I'll click on add new endpoint. We'll name the service. I'll just call it network check.
For the URL, I'll add my URL, networkcheck .com. And then we can specify settings by clicking on this button here, individual settings to use for this endpoint. Or we can check for status codes too, like 404 and things.
We can search for strings. So like I can search for in the body. So you can actually make sure your website is saying your stuff and maybe it didn't get like taken over or hacked.
And then just like before with Proxmox, under conditions, I'll say for an explicit host, networkchuck .com. Click on save. And we'll go back to our setup, back to our host.
Over here to network chuck, I'll do save and run service discovery. Rescan right now. And there it is.
The service is HTTPS network chuck. And we're using the HTTP plugin. Oh, I forgot to activate the changes.
All right. Now let's go to monitor and all host and network .com is up clicking on that. And if we click on that service, I think we get more info here.
Oh yeah. Look at all this expected string network. Chuck found it, found it.
All right. I do wanna change one thing. I'm gonna change the rule to check for a different word to see if it fails.
Let's say the string will be, I don't know, pickles. All right. Now like something real, like, uh, Monero, save that, commit that.
And if we go to monitoring, we should see something fail, right? Actually probably have to do our host discovery again. Oh yeah, there it goes.
Critical. Expected string not found. Monero not found.
But let me undo those changes. Now I told you we're gonna monitor everything and one thing I really want to monitor is my recording machine. The one I'm on right now.
This is my Windows machine. It's a gaming PC. Let's see how good Check MK's Windows monitoring is.
And I haven't seen this at all yet. Really excited to see that. So for windows, we'll go to set up and we'll go to agents and click on windows and we should see the agent right there.
I can click on that and actually download straight from there. Yes. Keep it.
And let me install it a law, accept it, install, click on finish. And then we'll add our host set up host at host. My machine's name is Deb too.
Comment below. If you remember Deb, do you know who Deb is? If you do, you're an OG.
I'll specify my IP address for this host. Let's test the agent connection. Oh, we have a firewall issue.
I need to add a new rule. Let me open up my terminal and add that rule right now. And I'll probably need to open this inside a privileged terminal.
So I'll right click that and say run as admin. I'll run it. New rule made.
Let's try the test one more time. Retest. Oh, it hates me.
Okay, the firewall rule didn't quite work. All I had to do was restart the check market agent service, check MK service on my PC, and it worked. And that was all I have to do.
Save and run. It's updating. And what did it find?
CPU, disk, file system. I'm running out of space. Dude, ton of stuff.
I want it all, accept all, apply all, activate everything. Okay, Deb is all ready. Let's click on Deb and there's all the stuff.
Now look at this. On my interface, my network interface, I've got some stuff, some warnings. Speed, it should be 10 gig.
It's five gig. That's some errors. I mean, this is already paying off file system, man, almost out of space.
My windows time service is not running. Why I'm learning some stuff right now. My offset is two seconds, which is like forever in time world.
Now I know what you might be thinking, Chuck, this is just monitoring software. We've seen this before. You've shown stuff like this before.
What else does this do? Let me show you. So like right now, this server, I told you it's a special server, but you can't see that from here.
because it's not monitoring everything just yet that I have on the server. The default plugins don't do it. So I'll need more plugins.
That's the secret. Now, speaking of stink and plugins, I just found something really bad. Watch this.
It's actually one node and a four node cluster. One, two, three, four. It's a C cluster and that's my NAS.
And I added the C plugin so it could see all the NAS stuff. And I found something nasty. Dude, look at my SSD pool.
97 % full. I have to offload this video recording right now to this. And I don't know how I'm gonna do that.
I gotta fix it. This stuff is already helping me while I'm recording for this. So powerful.
Thankfully check MK. I said it right that time. Yes.
You owe me a sip of coffee. That's the drinking game we play. Thankfully they do have plugins for Seth and also notice we don't have any hard drive information either.
So in the same way we added it for Docker, we're gonna add plugins for my Seth and my hard drives, the smart POSIX stuff. And so with Seth and my hard drive stuff, I can see everything on every one of our hard drives.
And there's a lot of hard drives in there. Look at this. Isn't this awesome temperature and everything.
Now, if you're new to monitoring software, getting all of this like this is not normal. Normally you're very limited by the number of hosts and services you can monitor. And this gives you everything.
This is the ultimate home lab thing. It's the ultimate business thing. I'm a business and I'm running this for me.
It's freaking great. Being able to see every stinking little thing I do and monitor it and oh, just, oh, I love it so much. Now I want you to get back to this idea that you can literally monitor anything with this software.
So far we've been doing a bunch of built in things, things you might expect from a monitoring software, but you know what felt kind of lacking the unify stuff. I wanna see more. And thankfully you can find a plugin for that.
What do you mean find it's not here in check MK. They actually have a marketplace where people can submit their own plugins and there's one for unify. You wanna see it?
Let's go. I haven't tried this yet, so I'm really excited. The site is exchange .checkmk .com.
I'm gonna browse some packages here. Search for Unifi. There it is right there.
Unifi controller special agent. Now, of course, install at your own risk, but I'm feeling dangerous. I'll switch over from the description to all versions right here, and I should see my file.
I can download it, but what I think I'll do is I'll just right -click and copy that link, and I'll get logged into my Checkmk server. I'll download this sucker, curl -O. paste that in.
Got it. And then I'll do MKP the command MKP add unify dash three dot zero dot zero dot MKP. It's a file we just downloaded.
See what happens. Oh, command MKP not found. Oh, I think I have to become the user.
So I'll do pseudo SU SU is for switch user and I'm gonna become the monitoring user. Is that what I have to do? Yes.
That's what I had to do. Okay. Now let's do that same command.
Look at us learning together. Let's run it. Add yes.
And now we have to enable MKP enable. unify -3 .0 .0 .mkp. Oh, wait, wait, wait.
No, no, no. You don't enable the file name. You say enable unify -3 .0 .0.
Okay, cool. So we downloaded a plugin from the marketplace. We installed it, we enabled it.
So then we'll get in there and add the rule, add a few parameters, do service discovery. I also had to do that weird folder thing to import all the hosts that the controller might find. And boom, now I have stuff.
Now look at my host, for example, my living room AP. I get all the info and that's all through piggybacking off that one API service. And I can see like all the SSIDs and their performance right from here.
That's pretty sick. Okay, cool. We found a plugin and we installed it, but what about stuff we wanna monitor?
That's not out there. What if we want to build a plugin and we can do that? Plex is something I want to add.
I wanna see everything with my Plex server. I wanna monitor everything in one place. Is that possible?
Yes, it is. So I'm gonna have my AI build a plugin for me. Here's what I wanna see.
I wanna see monitoring status. I wanna see who's watching, who's watching what, how many shows. Build the thing.
Here's my Plex token. It's already on the server. Throw the plugin on there and test it for me.
Go. Get to work. Coffee break.
I'm not kidding. That built it wicked fast. Now I need to add that rule and the host.
Set up, host, add host. You know the drill. Now when I add the rule, I wanna look for integrate Nagios plugins.
There we go. I'll add a rule. I'll name the service Plex.
It'll be a command line run. Look at my Plex host. And it's really that simple.
Like we're just executing a script and telling check MK to look at that and make it happen. Safe. We'll activate.
Okay. Plex is ready to go. Check this out.
I'll get into that service actually right here. You see that I'm watching Seinfeld. It's got the number of everything.
It's so cool. How sick is that? And while you weren't looking, I added this to home assistant.
I got full home assistant monitoring here. I can see the number of cameras that are. up and that are unavailable, which is a lot of them.
What's going on? The number of entities over time. And then I added this to air quality, number of lights that are on and off.
I see it changing right there. Motion events. Look at that.
And this is just like, hey, I want to see this. I could add so much more and monitor everything. Now, by the way.
what these are called. These checks we're doing are called active checks. They're not installed on the device itself.
Like the thing that we're monitoring, like home assistant, there's nothing running on there. It's a script that we have on check MK and it's going out and doing something. This one's probably the best thing I'm monitoring out of this entire video.
I'm not kidding. I'm monitoring my Hermes agents. Check this out.
I've got Hermes agent. Her name is Hermione. Here she is.
She's on her own VM. So when we jump in there, you're gonna see lots of things like her CPU load and CPU load. and file system, but you know what else you can see?
Her Hermes Crohn's. Are those still running? Cool.
Her Hermes Gateway, is that up? Yes. Her Hermes model login, it's good for 91 hours.
She's good right now. Is she replying? Has she been used?
I can see how much she's actually doing. These are just things I made up out of thin air. How about her usage?
How sick is that? If you're like me and you're running a lot of Hermes agents, being able to put them all in one interface and be able to monitor all this, are you kidding me? It's the best.
Now, by the way, the way I did this. She's a Linux machine like any other machine we have. So she hasn't check MK agent running on her, but you can actually have local scripts running that the agent will itself execute.
If I LS this directory here, this is where we put those scripts and have one called Hermes agent. If I run this, this is all it's doing just this. And it's reporting back as a metric to check MK.
So whatever you can program or script right here, you can monitor anything, anything, and just putting it there shows up as a service on Hermione and check MK. Okay. Monitoring all this stuff is great, but what happens when things.
go down. How will we find out? I can just sit here and watch this all day, even though I kind of want to, we need alerts and we can do that.
Absolutely. I'm gonna show you real quick what it looks like to set up an alert with slack. I'll go to set up notifications over here under events.
And at the top here, I'll click add notification rule. We're gonna keep the defaults next step filter, keep as defaults. Now our method won't be email.
It's gonna be slack and you can see we have other options as well. Like matter. Most we'll create our parameters.
It's gonna be a web hook and I have this in slack over here. I'm not gonna show it to you though. I'll save that next.
The recipient will add a recipient CMK admin next, next, next apply and test notification rule. Actually, I won't test it. I wanna, I just wanna do it live.
So, uh, I'm gonna break Hermione. I'm gonna stop her gateway. Her gateway is shutting down.
The monitor should pick it up after about a minute and I just got it. If you look right here, there it is sent to Slack. And then I just did that live with you.
Like that was really easy to set up. I need to bring her back. Cuz people are gonna get mad.
Actually should send me a message when she comes back up. Boom recovery. She's back up.
I love it. This is so sick. Right now.
What if I had an issue and then it got sent to my AI IT department and they solved it. Let's test that. So set up notifications.
I'm going to focus on the filtering of host. And I'm just going to look at. And for her service filter, I want it to be just the Hermes gateway.
And then my notification method is gonna be a plugin, a script that I already installed called Paperclip, which I just made a video about that. Check it out somewhere around here. It's how I run my IT department.
Essentially, this is just telling it how to reach Paperclip and create a task on Paperclip. I'll create some parameters, really just the description, that's it. Next for recipient, we're moving all contacts.
Don't want that. I'll just have it go to the default CMK admin. And lastly, with general properties, I'll just call this ticket to the IT department.
Let's review it and we'll apply. Now I'm going to open up a paperclip. I'm going to bring Hermione down.
Her gateway is stopped and we'll see if a ticket comes in for Ron to fix it. Yes, are you seeing this? This is so freaking cool.
Let's see what the ticket says. All right, let's see what he does. Okay, Ron just finished.
Let's see what he did. I'll go to my decisions and he did it. He restarted the gateway.
Let me go check and see if the work was done. Let's see if Hermione is indeed up. Yeah, running.
It's been up six minutes. Good job, Ron. How cool is that?
IT department, man, we got it going. Now, time for probably the best part of this, and that's being able to hand over the keys to this monitoring software, CheckMK, to an AI agent and say, hey, do this for me. Like, once you've gotten the ropes like you did in this video, now you can feel confident to go, you know what, I know how this works.
Let me have my AI agent do this for me. Like, I added some of my hosts. but I have a lot more to add.
I just wanna monitor things. So anyways, check this out. We're gonna go to setup.
We're gonna go to, I'm just gonna search for users. There we go, users. And then we'll go to add user.
Name it, I don't know. Maybe I'll give Ron access to this. Ron Weasley.
And the key here is we're gonna make it an automation seeker for machine accounts. And for the roles, we'll make him an administrator. And that's it, save.
And then I'll commit Ron's changes, activate, and that's it. Now you might notice that I already have an automation in there for Morpheus. Morpheus is my main AI.
And throughout this video, he has helped me with a few things. Like when I didn't show adding scripts, Morpheus was doing that. And now I'm gonna do two things, actually.
I'm going to tell Morpheus to add the rest of my hosts. I'm gonna tell him right now. I'll also say create a beautiful dashboard and check mark.
Check MK for me as well. AI is gonna take care of this for us. I'll see you on the other side of that.
Uh, he finished. And I have every single host. Look at this.
So many stinking things. Oh, it's great. My YouTube warehouse, which has two critical things.
What's going on there? Snapshot age, system D service. Okay, but you can see it all.
Like if I go to monitor and I go to see Docker containers, there they all are. Oh, isn't it beautiful? And then check this out.
I had to make a custom dashboard. If I go to monitor, it should be, oh, right here, the studio. Look at that.
It looks nice. You can do whatever you wanna do. That's what's crazy with this.
I can look at host and service problems, get a glance at that. I can look at service problems by themselves, look at history, downtime history. We don't have any cuz all my stuff is always up.
I assume I never knew, but now I have visibility into my environment. This is so exciting and I can monitor little, literally anything I, I want with any service, with any VM, with any AI agent, it's all done. I wanted to have.
a Hermes dashboard created for me. So I'll go to monitor and then Hermes agents. And this is all of my, oh, and look at this.
George and Luna are critical. Why? No backup found for Luna.
I'm seeing some login issues for models. I mean, this is pretty sick. Monitoring all my Hermes agents at a glance.
So that is Check MK, Check Mark, whatever you want to call it. It's awesome. You can monitor whatever you want.
Let me know in the comments what you're going to monitor. I know this video is probably long. I kind of see this video as like a a large tutorial and you can jump in at any point and learn how to do one thing.
But if you stay till the end, you're a real one. I appreciate you. Thanks again to check Mark.
Dang it. I said it again, check MK for sponsoring this video. I love sponsors that just want me to talk about their free tools.
Sure. There is something paid on the other side of this. And if you are a legit enterprise, this is gonna work for you.
But for us home labbers and small businesses, I just wanna have really great monitoring for all our stuff. This solution's crazy. Oh, and here's something cool.
Check MK also has a free trial of their most advanced commercial edition check MK ultimate. And that trial stays free for up to 750 services on one site. That's a lot of monitoring for $0 and being able to connect your AI agents to this and just have your AI agents build this out for you.
Oh, so good. And also you gotta hack that YouTube algorithm today. Hit that like button, subscribe, comment notification bell.
You gotta hack YouTube today. Ethically. Of course.
That's all I got. I'll catch you guys next time. Hey, you're still here.
If you're new here, I like to pray for you at the end of my videos. It's a little weird. I know.
Just go for it. There's plenty of things in life that are weird, like my beard, hairs trying to curl up on me. My mustache hairs, rather.
So I'm just going to pray for you. Ready? One, two, three, pray.
God, I thank you for this person that I'm talking to right now. I thank you for who they are. They're wonderful.
human being. They're fearfully and wonderfully made. I thank you for their talent and their humility and just the gift of their life.
I thank you that they're watching this video and they're learning about monitoring their IT stuff. But God, it comes to mind that we tend to monitor things that are nerdy, that we care about, that might go down, that kind of feed our hobbies, but we often don't monitor the things in our life that need attention, like our families and our health and our spiritual well -being.
So Lord, I'm just asking your name that you would give, actually give me and the person I'm praying for right now, just the insight into our lives of like, what are we missing? What do we need to monitor? What do we need to pay attention to?
What in our lives, needs our attention because there's something that we're missing. There's a critical alert that we're not really paying attention to.
Is it with our kids? Is it with our spouse? Is it with our friends and our family?
Like, what is it? God, just give us that ability right now. Show us a way.
And just bless this person on the other side of the screen, on the other side of this camera. Just bless their life and what they're doing. Go before them right now and make their path straight.
God, you say we plan our ways. We plan. We plan a lot, but you direct our steps, and I pray you do that for us today, God.
It's in Jesus' name I pray. Amen. Thank you for letting me do that.
I will catch you guys next time.
The Hook
The bait, then the rug-pull.
The cold open is a list of things that could be broken right now without you knowing: a Linux server, a gateway, a NAS, a hard drive about to fail, Docker, Proxmox, UniFi, Plex, Home Assistant. The promise that follows is unusually literal. By the end, every one of those is a row in one dashboard, and so are the AI agents, the Windows gaming PC, and a Plex server reporting who is currently watching Seinfeld.
Named ideas worth stealing.
03:40list
Five ways to collect data from a host
- Ping: is it alive, nothing more
- Installed agent: deep local data from inside the machine
- SNMP: standard status exposed by network gear, no install needed
- Special agent: the monitoring server queries the target's own API
- Piggyback: one host reports on behalf of children that have no IP
- Active and local checks: run your own script, from the server or on the host
Every section of the video is really the same decision made differently. Before adding anything, you pick which of these the target can support, and that choice determines whether you install software, open a port, write a rule, or write a script.
Steal forany monitoring or integration build where you have to decide between installing something, calling an API, and scraping a parent system
05:00model
The add-a-host loop
- Add the host with a name or IP
- Choose the collection method under monitoring agents
- Test the connection before you go further
- Save and run service discovery
- Monitor undecided services, accept all
- Activate pending changes, then wait about five minutes for data
The same six steps repeat for a Linux box, a switch, a container, a VM, a website, and a Windows PC. Once you have seen it twice the rest of the video is pattern matching.
Steal forteaching any tool with a repeating configuration loop, where the loop is the lesson and the variations are the examples
14:40concept
Piggyback for things without an address
Docker containers have no IP you can reach, so you configure a folder with no IP, no agent, and no API, set it to always expect piggyback data, and import the container names. The parent Docker host's agent supplies the data and Checkmk splits it across those named host objects.
Steal formodelling child objects in any system where the children cannot be contacted directly
17:30concept
Agent and API together, not either-or
The installed agent is described as an inside man reporting from within the machine, while the special agent knocks on the API from outside. Running both against the same host is explicitly recommended because they see different things.
Steal forany data collection problem where an internal and an external source are assumed to be alternatives
29:50concept
Anything scriptable is monitorable
A script in the agent's local checks directory, or a command line wired through a Nagios plugin rule, becomes a service like any built-in check. This is how Plex watch sessions, Home Assistant entity counts, and AI agent heartbeats end up in the same dashboard as CPU and disk.
Steal forextending any tool past its documented integrations instead of waiting for official support
How they asked for the click.
VERBAL ASK
35:50subscribe
“Let me know in the comments what you're going to monitor. Hit that like button, subscribe, comment notification bell.”
The sponsor is named up front at 1:05 and again at the end, with the unusual framing that the sponsor only asked him to show a free tool. The comment prompt is specific to the video's topic rather than a generic ask, and there is a second nested hook, a reference to an earlier AI IT department video, placed right at the alerting payoff where interest peaks.
MENTIONED ON CAMERA
FROM THE DESCRIPTION
PRIMARY CTAWhere the creator wants you to go next.
OTHER LINKSAlso linked in the description.
- Join the NetworkChuck Academy! ↗
- 📄 Follow-along guide (every command from the video) ↗
- 🧩 Checkmk Exchange (community plugins, including the UniFi one) ↗
- 📖 Checkmk documentation ↗
- 📖 Monitoring Docker (piggyback setup) ↗
- 📖 Monitoring Proxmox VE (special agent) ↗
- 📖 Local checks (how I monitor my Hermes agents) ↗
- 📖 Checkmk REST API (what my AI agents use) ↗
- 💰 Compare the Checkmk editions ↗
- 💻 Checkmk on GitHub ↗
- 📺 My AI IT department (Paperclip) ↗
- 📺 Set up your own AI agent (Hermes) ↗
- ☕☕ COFFEE and MERCH ↗
- 📚 CCNA Course ↗
- Instagram ↗
- X/Twitter ↗
- Facebook ↗
- Join the Discord server ↗
Visual structure at a glance.
