DNS_PROBE_FINISHED_BAD_CONFIG points at your own settings rather than at a website. Chrome tried to resolve a name, got nothing usable from the resolver it was configured to use, ran its own probe, and concluded the DNS configuration itself is at fault. It is generally right, and the fix is usually one setting.
Prove it is local in ten seconds
Open the same URL on a phone with Wi-Fi off, using mobile data. Different device, different resolver, different everything. If it loads there, the site is fine and this page is about your machine.
The fix that works most of the time
Point the device at a public resolver instead of whatever your router is handing out.
- Windows - Settings → Network & Internet → your adapter → Edit DNS server assignment → Manual → IPv4 on → 1.1.1.1 and 8.8.8.8.
- macOS - System Settings → Network → your connection → Details → DNS → add 1.1.1.1 and 8.8.8.8, remove the others.
- Android - Settings → Network & internet → Private DNS → Private DNS provider hostname →
one.one.one.one.
Then flush the caches, because a stale negative answer will otherwise survive the change:
ipconfig /flushdns # Windows
sudo dscacheutil -flushcache; sudo killall -HUP mDNSResponder # macOS
resolvectl flush-caches # LinuxClear Chrome's own cache too, at chrome://net-internals/#dns. Chrome caches separately from the operating system, which is why an OS-level flush on its own so often appears to change nothing.
If names resolve immediately after this, you are finished. The resolver your router was advertising was the whole problem - a very common state on consumer routers that have been up for months.
If that did not fix it
Check what is actually configured
Settings dialogs sometimes disagree with what the stack is really using, particularly when a VPN or security product has written its own values:
nslookup google.com # note which server it says it used
resolvectl status # Linux
ipconfig /all | findstr /i "DNS Servers" # Windows
scutil --dns | head -20 # macOSIf the server listed is not the one you set, something is overriding it. That is almost always a VPN client, a corporate management profile, or an antivirus with DNS filtering.
Rule out a VPN or proxy
A VPN that lost its connection without disconnecting cleanly leaves DNS pointing into a tunnel that is gone. Connect it and disconnect it properly rather than just quitting the client - that restores the original settings, where force-quitting does not.
Reset the network stack on Windows
Third-party software inserts filter drivers into Winsock and does not always remove them cleanly, which produces exactly this symptom after an uninstall or an update:
netsh winsock reset
netsh int ip reset
ipconfig /release
ipconfig /renewRun those from an elevated Command Prompt and reboot. They do not take effect until you do.
Check the router
If every device on the network shows this, the router is the common factor. Log in and look at its WAN DNS settings - a manually entered resolver that has since been retired is a common cause on routers configured years ago. Setting them to 1.1.1.1 and 8.8.8.8 at the router fixes every device at once. A firmware update is worth doing while you are in there.
The case where it is not your configuration
Occasionally an ISP's resolver fails for everyone on it at the same time, and every customer sees a DNS error while the rest of the internet works normally. The tell is that switching to a public resolver fixes it instantly on every device you try. That is not something you can fix, but it is worth recognising rather than resetting a network stack that was never broken.