Notes & guides
Diagnostic how-tos, HTTP reference, and notes on how isitdown.io probes the internet. Written for operators and curious users.
Anatomy of a website outage: how sites actually go down
What happens between a working site and a 503 error — the failure categories, real named outages from each, and why status pages always lag the real impact.
HTTP 502 Bad Gateway: what it actually means and how to fix it
What 502 means, how it differs from 503 and 504, the half-dozen things that actually cause it, and how to diagnose one as either a user or an operator.
HTTP 504 Gateway Timeout: what causes it and how to fix it
504 vs 502 vs 408 vs Cloudflare 524 — what each one means, what actually causes a 504 in production, and how to diagnose one fast.
DNS_PROBE_FINISHED_NXDOMAIN: what it means and 6 ways to fix it
What Chrome's NXDOMAIN error actually says, why it appears for sites that absolutely exist, and the diagnostic checklist that fixes it in under a minute.
SSL certificate expired: what it means and how to fix it
Why expired certs break sites in seconds, the calendar-driven outages this causes every year, and the operator + user playbook for fixing or working around one.
Cloudflare Error 520: 'Web Server Is Returning an Unknown Error'
What 520 actually says, the difference between 520 / 521 / 522 / 523 / 524, the 5 origin behaviors that cause it, and how to debug one with the cf-ray header.
Your Minecraft server says 0/0 players. Here's what's actually happening.
If your server pings as online but reports zero players when you know there are people on it, the cause is almost always one of four protocol-level quirks. Here's how to tell which.
How to check if a Minecraft server is online (without launching the game)
Hypixel, 2b2t, EarthMC, and the rest — a quick guide to telling whether a server is actually offline, just queueing you, or whether the problem is on your end.
A2S_INFO is the fastest smoke test for a Source-engine server
Before you SSH into the box, before you check the logs — fire one A2S_INFO packet. The response (or its absence) tells you which of five common failure modes you're looking at.
What runs on AWS? A field guide to the cascade
When AWS us-east-1 hiccups, dozens of unrelated consumer services go dark together. Here's the dependency map most people don't realize exists — plus how to read an outage when it actually happens.
Is it down for everyone or just me? How to tell in 60 seconds
A three-step diagnostic that tells you whether a site outage is global, regional, or just your own network — before you escalate anything.
What does HTTP 503 Service Unavailable actually mean?
The difference between 503 and 500, 502, 504 — what each code means, what causes 503 specifically, and how to diagnose one in under a minute.
Multi-region website checks: why a single ping isn't enough
A single-region ping can't tell a global outage from a regional one. Here's what probing from 4 regions in parallel reveals — with real examples.
Cloudflare Error 1020 Access Denied: what it means and how to fix it
Error 1020 is a deliberate block by the site owner's Cloudflare firewall, not an outage. Here's how to diagnose why you're blocked and what to do next.
DNS propagation: why it really takes hours (and when it doesn't)
DNS doesn't actually propagate — resolvers cache records until their TTL expires. Here's what controls the real wait time and how to speed it up.
Flush DNS cache: exact commands for Windows, macOS, Linux, and Chrome
The right command for every OS, when to use it, and how to verify the cache actually cleared. Plus what flushing won't fix.
HTTP 429 Too Many Requests: what triggers it and how to handle it
What 429 means, how it differs from 503 and 1015, when servers send it, and the right way to back off without making the problem worse.
Cloudflare Error 1015: what "You are being rate limited" actually means
1015 is a Cloudflare-edge rate limit, not an outage. Here's why it triggers, how it differs from 1020 and 429, and what to do as a user or site owner.