ERR_SSL_VERSION_OR_CIPHER_MISMATCH: how to fix it

"This site uses an unsupported protocol" means the browser and server share no TLS version or cipher. How to find out which, and fix it on nginx, Apache, IIS, Cloudflare and other hosts.

· 6 min read · By the Spot Downtime team

Chrome and Edge show ERR_SSL_VERSION_OR_CIPHER_MISMATCH with the message “This site can't provide a secure connection. example.com uses an unsupported protocol.” Firefox calls the same failure SSL_ERROR_NO_CYPHER_OVERLAP.

It means the HTTPS handshake started, but the browser and server couldn't agree on how to encrypt the connection: there was no TLS version, or no cipher suite, that both of them support. There's no “proceed anyway” button, because there's no secure connection to proceed with.

Visiting a site, not running it?

Update your browser and operating system, and check your device's date and time. Then try without a VPN, and with antivirus “HTTPS scanning” turned off. If the site still fails on an up-to-date browser, the problem is on the site's side, and only its owner can fix it.

The causes, most common first

  1. The server only offers TLS 1.0 or 1.1. All major browsers dropped them in 2020 (Chrome 84 and later). Old servers and appliances that never enabled TLS 1.2 now fail for every visitor.
  2. There's no certificate for this hostname at the edge. On Cloudflare and other CDNs, a hostname the edge has no certificate for can't complete a handshake. This is the most common cause on Cloudflare.
  3. The cipher list is obsolete or empty. A configuration copied from an old guide that only allows RC4, 3DES or other ciphers browsers removed years ago, or a typo that leaves no valid cipher at all.
  4. The certificate type doesn't match the ciphers. An ECDSA certificate with only RSA cipher suites enabled, or the other way around, leaves nothing to agree on.
  5. The client is too old. Windows XP, Android 4.x and other old systems can't speak the TLS versions modern servers require.

Find out which one it is

Test each TLS version against the server directly. Always pass -servername, so the server picks the right certificate:

bash
openssl s_client -connect example.com:443 -servername example.com -tls1_2 </dev/null
openssl s_client -connect example.com:443 -servername example.com -tls1_3 </dev/null
  • Both fail with handshake failure or no protocols available, but -tls1 works: the server is stuck on old TLS.
  • Both fail, and the server's IP belongs to a CDN: the edge has no certificate for this hostname.
  • TLS 1.2 fails only with modern ciphers: the cipher configuration is the problem.
  • Everything works here but not in one browser or on one device: look at the client, a proxy or antivirus.

To list every version and cipher the server accepts:

bash
nmap --script ssl-enum-ciphers -p 443 example.com
TLS version checkerSee which TLS versions a server accepts, the cipher it negotiates and whether its certificate is trusted, without installing anything.

Fix it on your server

The goal is the same everywhere: enable TLS 1.2 and 1.3, and use a current cipher list. Mozilla's SSL Configuration Generator produces tested settings for most servers; its “intermediate” profile works with every browser from the last several years.

nginx

nginx
server {
    listen 443 ssl;
    server_name example.com;

    ssl_certificate     /etc/ssl/example.com/fullchain.pem;
    ssl_certificate_key /etc/ssl/example.com/privkey.pem;

    ssl_protocols TLSv1.2 TLSv1.3;
    # Leave ssl_ciphers at nginx's default, or use Mozilla's intermediate list
    ssl_prefer_server_ciphers off;
}

Check every server block that listens on 443 for the same IP and port: the protocol settings of the default server can apply to the others. Run nginx -t, then reload.

Apache

apache
SSLProtocol -all +TLSv1.2 +TLSv1.3
SSLHonorCipherOrder off

Look for older SSLProtocol or SSLCipherSuite lines in other included files (for example ssl.conf) that override yours, then run apachectl configtest and reload.

IIS and Windows Server

On older Windows Server versions, TLS 1.2 may be disabled in the registry under HKLM\SYSTEM\CurrentControlSet\Control\SecurityProviders\SCHANNEL\Protocols. Enable TLS 1.2 for the server and restart. IIS Crypto is a free tool that sets protocols and cipher suites with a “Best Practices” template. Windows Server 2008 and earlier can't be fixed properly: put a modern proxy or CDN in front, or upgrade.

Java, Node.js and other app servers

  • Java: very old JDKs lack TLS 1.3 and modern ciphers. Update the JDK, and check jdk.tls.disabledAlgorithms in java.security and the connector's sslEnabledProtocols.
  • Node.js: a custom ciphers or minVersion/maxVersion option on the HTTPS server can leave nothing in common. Remove it to use Node's defaults.

Load balancers and appliances

If TLS ends at a load balancer, the setting lives there, not on your servers. Cloud load balancers use a named TLS or SSL policy: choose a current one that includes TLS 1.2 and 1.3. Old hardware appliances may need a firmware update.

Fix it on Cloudflare

Cloudflare's free Universal SSL certificate covers your apex domain and one level of subdomains: example.com and *.example.com. It does not cover deeper names like dev.api.example.com, which then fail with this error. To fix it:

  • Check the edge certificate in SSL/TLS → Edge Certificates. It should be “Active” and list your hostname.
  • Multi-level subdomains need an advanced certificate (Advanced Certificate Manager) that includes the name, or a shallower hostname such as dev-api.example.com.
  • New domains or just-proxied records can show the error for a while until the certificate is issued. Wait, then check again.
  • Disabled Universal SSL leaves no certificate at all unless you uploaded or ordered one.

Other CDNs and hosting platforms work similarly: the hostname must be added to the platform and have an issued certificate before HTTPS works on it.

After the fix

  1. Run the openssl commands again: TLS 1.2 and 1.3 should both connect.
  2. Load the site in a private window, so the browser doesn't reuse a cached failure.
  3. Test from a phone on mobile data, to rule out your own network.

Catch it before your visitors do

This error stops every HTTPS visitor at once, and it tends to appear after a change nobody connected to the site: a new subdomain, a load balancer policy update, a server migration or a copied config. A website monitor makes a real HTTPS request on every check, so a failed handshake raises an alert within minutes, and SSL monitoring warns you before certificates expire.

Related: ERR_SSL_PROTOCOL_ERROR, when the server doesn't speak TLS at all, and Cloudflare error 525, the same handshake failure between Cloudflare and your server.

Get new guides by email

Practical posts on monitoring, outages, error fixes and on-call, when they're published. No spam.

Unsubscribe from any email with one click. See our privacy policy.

Keep reading