Outage Status Now

Guides / Chrome ERR_CONNECTION_TIMED_OUT / DNS_PROBE at work

Part of: Network, VPN and 'is it just me?' problems

Last reviewed · By

Chrome ERR_CONNECTION_TIMED_OUT / DNS_PROBE at work

Chrome ERR_CONNECTION_TIMED_OUT or DNS_PROBE_FINISHED_NO_INTERNET at the office? Split ISP, DNS, proxy, VPN and a site that is actually down.

Chrome shows ERR_CONNECTION_TIMED_OUT, DNS_PROBE_FINISHED_NO_INTERNET, DNS_PROBE_FINISHED_NXDOMAIN, or ERR_NAME_NOT_RESOLVED. The user is on office Wi-Fi. Tickets arrive as "the internet is down" or "Microsoft is down" depending on which tab they noticed first. You need a 90-second split.

The 90-second split

  1. Phone on LTE. Open the same URL off Wi-Fi. Works on LTE, fails on office Wi-Fi → office path (DNS, proxy, firewall, captive portal). Fails everywhere → destination or a wider ISP issue.
  2. Another site. Can they reach outagestatusnow.com, a newspaper, and their bank? All fail → local network. Only one hostname fails → that service or a targeted block.
  3. Status for the destination. If the tab is Microsoft 365, AWS console, or a Cloudflare-fronted shop, open our matching status page before you rebuild Chrome.

What the Chrome codes usually mean

Office network checklist

  1. Captive portal. Guest Wi-Fi that never finished login produces NO_INTERNET probes. Open a never-cached HTTP site to trigger the portal.
  2. DNS. nslookup / Resolve-DnsName against the corporate resolver and against 1.1.1.1 (if policy allows). Corporate DNS outages look exactly like "the whole internet died" while ICMP to the gateway still works.
  3. Proxy settings. Pac file unreachable → every HTTPS tab times out. Check system proxy and whether Chrome is using it (chrome://net-internals/#proxy).
  4. VPN. Full-tunnel VPN to a dead concentrator = timeouts to everything. Disconnect VPN and retest; see VPN connected but no internet.
  5. Firewall category / geo block. Newly aggressive web filters time out instead of showing a block page. Test with a known-allowed destination.
  6. IPv6 half-break. Broken IPv6 with Happy Eyeballs misbehaviour causes long hangs. Temporary IPv6 disable is a diagnostic, not a permanent fix.
  7. MTU / fragmented paths after a new SD-WAN — less common, but classic after VPN MTU mistakes.

Single-PC causes

When the destination really is down

If LTE and office both time out to one hostname, run our URL checker, check the vendor status page, and look for Cloudflare 5xx branded pages (owner-side origin faults look different — see Cloudflare 521/522). Do not change office DNS to "fix" a vendor outage.

When to escalate

Escalate to network engineering when many users on a site hit timeouts, DNS logs show resolver failure or firewall drops, and LTE works. Escalate to the SaaS vendor only when the hostname fails from multiple networks and their status or your multi-region probes agree. Include failed URLs, Chrome error codes, UTC times, and whether DNS resolved.

Related status pages

Related guides

FAQ

What does ERR_CONNECTION_TIMED_OUT mean in Chrome?
Chrome gave up waiting for a TCP connection. Causes include office firewall drops, dead Wi-Fi/VPN paths, broken proxies, or a server that is not answering.
What does DNS_PROBE_FINISHED_NO_INTERNET mean?
Chrome believes the network cannot resolve or reach the internet. Check Wi-Fi login portals, DNS, VPN and whether other devices on the same SSID work.
How do I tell an office network fault from a website outage?
Try the same site on mobile data and try other sites on office Wi-Fi. Office-only failure is local; all-network failure to one host is the destination.
Should I change DNS to 8.8.8.8 at work?
Only if IT allows it as a diagnostic. Many offices require internal DNS for split-horizon names; bypassing it can break intranet access.

Did this help?

Still broken? Check another tool