← Harden & defendsee everything your server does, live, for $0 with netdata

see everything your server does, live, for $0 with netdata

$docker run -d --name=netdata netdata/netdata

your server is talking. are you listening?

most people find out their server got popped when something breaks, a bill spikes, or a friend texts "hey your site is sending me weird links." by then the problem has been running for days or weeks. security problems almost always show up as weird metrics first, a cpu that's pinned for no reason, a network card pushing traffic at 3am, a process eating memory like it's getting paid for it. if nobody's watching those numbers, nobody notices until it's a mess. that's the whole pitch for netdata: free, fast, and it shows you what "normal" looks like so you can spot "not normal" immediately.

what netdata actually is

netdata is a monitoring tool that runs on your box and collects per-second metrics, cpu, ram, disk io, network throughput, every running process, and a few thousand other things depending on what's installed. it builds a live dashboard in your browser with zero config required to get started. no agents to babysit, no cloud account needed, no cost. it also ships with sane default alerts, so if something spikes outside the normal range it flags it instead of making you stare at graphs all day.

the command, broken down

docker run -d --name=netdata netdata/netdata

docker run starts a new container. -d runs it detached, meaning it stays running in the background instead of tying up your terminal. --name=netdata just gives the container a friendly name so you can reference it later with docker stop netdata or docker logs netdata instead of some random hash. netdata/netdata is the official image pulled from docker hub.

for this to actually collect real host-level metrics instead of just container stats, you'll want to run it with more flags for host networking and mounted system paths. the one-liner above gets you a working demo fast, but check netdata's docs for the full production run command with volume mounts for /proc, /sys, and /var/run/docker.sock if you want full visibility.

what you're actually watching for

once it's running, you'll see dashboards update every second. here's what matters for defense specifically:

cpu that's pinned at 100% with no obvious reason. cryptominers love doing this quietly in the background. if a process you don't recognize is eating your cpu 24/7, that's your first clue.

outbound network spikes at weird hours. your server doesn't usually push gigabytes of traffic at 3am. if it suddenly does, something might be exfiltrating data or your box got roped into a botnet.

disk io that doesn't match your workload. ransomware encrypting files in bulk shows up as a massive, sustained spike in disk writes. catching that early is the difference between losing a folder and losing everything.

new processes you don't recognize. netdata's process view shows everything running in real time. if you see something with a random name eating resources, that's worth investigating immediately, not next week.

lock the dashboard down

here's the part people skip and then regret. netdata's dashboard by default listens on port 19999 and shows a ton of detail about your system, running processes, software versions, network info. that's a gift to anyone scanning the internet if you expose it publicly. do not put this on the open internet.

firewall it so only your lan or vpn can reach it. a quick example with ufw:

sudo ufw allow from 192.168.1.0/24 to any port 19999
sudo ufw deny 19999

swap the subnet for your actual lan range, or better, only allow access through a vpn like wireguard or tailscale. that way the dashboard exists, but only you can see it.

the takeaway

monitoring isn't glamorous but it's the difference between catching a problem in minutes versus finding out from a stranger. netdata gives you that visibility for free, in one container, with alerts baked in so you're not babysitting graphs all day. spin it up, let it learn your normal, and firewall it off from the world. if you know someone running a server with zero monitoring, send them this. blind servers are the ones that get owned quietly.

watch the reel ↗
the weekly drop

one command a week that makes you harder to hack.

a single tool, explained in plain english, every week. straight to your inbox.

no spam. one email a week. unsubscribe anytime.