ERR_CONNECTION_TIMED_OUT & ERR_TIMED_OUT: Causes and Fixes

What Is ERR_CONNECTION_TIMED_OUT?

ERR_CONNECTION_TIMED_OUT is the error Chrome and other Chromium browsers show when they can't get any answer from a website's server. The page reads "This site can't be reached. example.com took too long to respond." Internally it's Chromium net error -118: "A connection attempt timed out" (net_error_list.h).

To open a connection, your computer sends the server a TCP connection request, a SYN packet, and waits for a reply. Normally the reply arrives within milliseconds. When no reply comes back, the operating system resends the SYN a few more times with growing gaps and then gives up; how long that takes depends on the system (Chromium's own notes put it anywhere from 20 seconds to 190 seconds). On our Mac, a connection attempt to an address where nothing answers gave up after 75 seconds.

Silence is the key detail. A server that is up but has nothing listening answers immediately with a refusal. With a timeout, nobody answered at all: the packets were dropped somewhere, or they were sent to an address where nothing exists.

ERR_CONNECTION_TIMED_OUT vs ERR_TIMED_OUT

Chrome has two similar-looking timeout errors, and they fail at different stages:

• ERR_CONNECTION_TIMED_OUT (-118): the connection was never established, because the server didn't answer the initial SYN. Look at the network path, firewalls, DNS and whether the server is online.
• ERR_TIMED_OUT (-7): a more general timeout from one of Chrome's own timers rather than the operating system's. The HTTPS handshake, for example, gets 30 seconds once the TCP connection is open (Chromium), and a connection that opens but then stops getting replies can end the same way. That usually points to a slow or stuck server, or a proxy that stalls.

The fixes below apply to both. ERR_TIMED_OUT more often turns out to be the website's problem, while ERR_CONNECTION_TIMED_OUT is more often a network path or firewall problem.

What Causes ERR_CONNECTION_TIMED_OUT?

Fix 1: Check Whether the Site Is Up for Everyone

If the server is down, nothing on your side will help, so check this first:

• Open the site on your phone using mobile data (Wi-Fi off). If it loads, the problem is your device or your network.
• Check it with the Website Down Checker, which opens the page from 11 places at once. If they all time out too, it's the site.
• Run a Ping Test from 11 cities. It times a TCP connection to port 443 from each city and adds a classic ping from our server. Answers on port 443 mean the web server accepts connections; if only the classic ping answers, the machine is up but its web port isn't.
• Test the web port with the Port Checker. Check port 443 (HTTPS) and 80 (HTTP). "Open" for us while your browser times out means something between you and the site is dropping the traffic. "Filtered" (no reply within 4 seconds) for us too means the server's firewall or the server itself is the problem.
• Run a traceroute from 8 cities over TCP 443 to see how far packets get. If every TCP 443 trace dies at the last hops, the site's hosting network is dropping the traffic; with ICMP probes, silent last hops are normal for many working sites.

Fix 2: Flush Your DNS Cache (and Chrome's)

When a site moves to a new server, your computer can keep using the old IP address until its cache expires. If the old server is switched off or firewalled, every attempt times out. Flush the operating system cache and Chrome's own host cache:

# Windows (Command Prompt)
ipconfig /flushdns

# macOS
sudo dscacheutil -flushcache; sudo killall -HUP mDNSResponder

# Linux with systemd-resolved
resolvectl flush-caches

# Chrome: open chrome://net-internals/#dns and click "Clear host cache"

To check whether your DNS resolver serves an outdated record, compare nslookup example.com (it asks your resolver directly, skipping the local cache and hosts file) with DNS Robot's DNS Lookup; sites behind a CDN can legitimately return different addresses. The flush DNS guide covers every browser and operating system.

Fix 3: Check the Hosts File

The hosts file overrides DNS for any name listed in it. Developers, ad blockers and some malware add entries, and a leftover line that maps the domain to an old or private IP will time out every time, on that one computer only.

• Windows: open Notepad as administrator, then open C:\Windows\System32\drivers\etc\hosts.
• macOS / Linux: run sudo nano /etc/hosts in Terminal.

Look for a line containing the domain that fails. Delete it, or add # at the start to disable it, save, and flush DNS again.

Fix 4: Turn Off Proxy, VPN and Firewall as a Test

If every site times out, a proxy or VPN that has stopped working is a common cause, because all traffic is sent to a dead end.

• Windows 11 proxy: Settings → Network & internet → Proxy. Under Manual proxy setup, select Edit (or Set up) next to Use a proxy server, turn Use a proxy server off and select Save. If Use setup script is on under Automatic proxy setup and you didn't set it, turn it off too.
• macOS proxy: System Settings → Network → your connection → Details… → Proxies → turn them off.
• VPN: disconnect completely and reload. If the site works without the VPN, change the VPN server or contact the provider.
• Firewall / antivirus: pause the third-party firewall briefly and reload. If that fixes it, add an exception for the browser instead of leaving the firewall off.

Fix 5: Switch to a Public DNS Resolver

Some ISP resolvers are slow to update, and some send blocked domains to an address that never answers. Switching to a public resolver such as Cloudflare (1.1.1.1), Google (8.8.8.8) or Quad9 (9.9.9.9) rules that out. Our DNS Speed Test shows which is fastest from your location.

If your DNS server itself is failing, you may see different errors, covered in DNS server not responding and DNS server unavailable.

Fix 6: Reset the Network Stack and Restart the Router

Restart the router and modem: unplug them for 30 seconds, then give them two or three minutes to reconnect. On Windows, also reset the TCP/IP stack from an administrator Command Prompt, then restart the PC to finish the reset:

netsh winsock reset
netsh int ip reset
ipconfig /release
ipconfig /renew
ipconfig /flushdns

On a Mac, turn Wi-Fi off and on, or remove and re-add the network under System Settings → Network. You can also run Windows Network Diagnostics (the link on Chrome's error page) and let Windows repair the adapter.

Fix 7: Test the Port From Your Own Computer

This tells you in under half a minute whether your machine can open a connection to the site at all, without the browser involved:

# Windows PowerShell
Test-NetConnection example.com -Port 443
# TcpTestSucceeded : False  -> no connection from your network (dropped or refused)

# macOS (-G sets the connect timeout)
nc -vz -G 5 example.com 443
# Linux
nc -vz -w 5 example.com 443
# "timed out" -> dropped; "Connection refused" -> server rejected it

If the test fails from your computer but DNS Robot's Port Checker shows the port open, something between you and the site is blocking it: your router, workplace firewall, ISP, a hotel or school Wi-Fi filter, or the site's own firewall blocking your IP address or country. Try another network, or ask the network administrator. (Test-NetConnection is documented by Microsoft.)

Fix ERR_CONNECTION_TIMED_OUT on Android and iPhone

• Switch networks: try mobile data instead of Wi-Fi, or the other way round. Captive portals on public Wi-Fi time out until you accept their login page.
• Android Private DNS: Settings → Network & internet → Private DNS → Automatic. A custom hostname that can't be reached breaks every lookup. See our Private DNS guide.
• Turn off VPN and ad-blocking apps, since many run as a local VPN.
• Reset network settings: on iPhone, Settings → General → Transfer or Reset iPhone → Reset → Reset Network Settings. On Android, Settings → System → Reset options → Reset Bluetooth & Wi-Fi (and Reset Mobile Network Settings if mobile data also fails).

For Website Owners: Why Visitors Time Out

If your site times out for everyone, the packets are being dropped before they reach your web server. Check these in order:

• Is anything listening? On the server, sudo ss -tlnp | grep -E ':(80|443) ' should show nginx, Apache or your app. If it shows nothing, start the service. Note that a stopped service usually produces refused rather than timed out, unless a firewall drops the packets first.
• Does the firewall allow 80 and 443? Run sudo ufw status verbose on Ubuntu or sudo firewall-cmd --list-all on RHEL-family systems. ufw's default deny drops packets, which is exactly what produces timeouts; firewalld's default zone rejects them instead, which visitors see as a refusal or an unreachable address.
• Cloud firewalls are separate. AWS Security Groups, Google Cloud firewall rules, Azure NSGs and many VPS providers' panel firewalls deny inbound traffic by default. Allow TCP 80 and 443 from 0.0.0.0/0 (and ::/0 for IPv6).
• Does DNS point to this server? After a migration, check the A and AAAA records with DNS Lookup. An AAAA record pointing to an address the server doesn't actually serve on can cause slow loads or time-outs for visitors on IPv6 networks.
• Is the server overloaded? If the connection backlog fills up under heavy load, new connection attempts are dropped and time out. Check CPU, memory and connection counts while it happens.

# Ubuntu: allow web traffic through ufw
sudo ufw allow 80/tcp
sudo ufw allow 443/tcp
sudo ufw status verbose

# Confirm from outside (another machine; on a Mac use -G 5 instead of -w 5):
nc -vz -w 5 yourdomain.com 443

Timed Out vs Refused vs Reset: How to Tell Them Apart

A firewall set to drop traffic produces a timeout; one set to reject it answers at once, usually as a refusal, or as ERR_ADDRESS_UNREACHABLE when it rejects with an ICMP message, as firewalld does by default. Detailed guides: ERR_CONNECTION_REFUSED, ERR_CONNECTION_RESET and ERR_NAME_NOT_RESOLVED.