Cloudflare Error 1033 (Argo Tunnel Error): Why Your Tunnel Can't Be Reached and How to Fix It
Cloudflare error 1033 means your hostname points at a Cloudflare Tunnel that isn't connected, or at the wrong tunnel. How to check cloudflared, status and DNS.
Cloudflare error 1033 means your hostname is routed to a Cloudflare Tunnel, but Cloudflare can't find a connected tunnel to send the request to. Usually cloudflared on your server has stopped or never connected. If the tunnel shows as healthy, the DNS record is pointing at the wrong tunnel ID, for example one you deleted and recreated. Your app itself may be running fine.
Checked against Cloudflare's Error 1033 page, the Cloudflare Tunnel troubleshooting, DNS, firewall and logging docs, and the cloudflared source in November 2026.
The error page shows something like:
Error 1033
Argo Tunnel error
Argo Tunnel is the product's old name. Cloudflare's FAQ explains it was renamed Cloudflare Tunnel once it stopped requiring a paid Argo subscription (in 2021), and the current Error 1033 page calls it "Cloudflare Tunnel error". Same error, same causes.
How a request reaches a tunnel
If you're new to tunnels, Cloudflare Tunnel Explained covers the setup. The short version of the path:
app.example.com → CNAME <UUID>.cfargotunnel.com → tunnel <UUID> → cloudflared on your server → localhost:3000
- Your DNS record for
app.example.comis a CNAME to<UUID>.cfargotunnel.com, where the UUID is the tunnel's ID. - Cloudflare looks up that tunnel.
- The tunnel only works if at least one
cloudflaredprocess (a "connector") is connected to Cloudflare. cloudflaredforwards the request to your app.
Error 1033 means the chain broke at step 2 or 3. If it breaks at step 4 (tunnel connected, app not answering), you get a 502 with "Unable to reach the origin service" instead. That's a different problem: the service URL, port or protocol in your tunnel config is wrong, or the app is down.
Step 1: check the tunnel's status
In the Cloudflare dashboard, go to Networking > Tunnels and find your tunnel. Or, from a machine where you've run cloudflared tunnel login:
cloudflared tunnel list
That lists your tunnels with their IDs and current connections. For details on the connectors of one tunnel:
cloudflared tunnel info myapp
The status tells you what to do next. From Cloudflare's troubleshooting page:
| Status | Meaning | What to do |
|---|---|---|
| Healthy | Connected and serving traffic through four connections | Tunnel is fine; go to Step 3 (DNS) |
| Inactive | Created, but cloudflared has never connected |
Install and run cloudflared on the server |
| Down | Was connected, but the cloudflared process stopped |
Find out why it stopped and start it (Step 2) |
| Degraded | Serving traffic, but at least one connection failed | Read the cloudflared logs and check the firewall |
Cloudflare also warns that the status only describes the link between cloudflared and Cloudflare. A tunnel can be Healthy while your app behind it is broken, which is the 502 case above.
Step 2: get cloudflared running again
On the server that should be running the tunnel, check the service. If you installed it with cloudflared service install, the systemd service is called cloudflared:
sudo systemctl status cloudflared
sudo journalctl -u cloudflared -n 100 --no-pager
(What is systemd? explains both commands.)
Common reasons it isn't running:
The machine is off or asleep. A tunnel running on a laptop disappears when the lid closes. Cloudflare's docs list power loss, application crashes and network changes as the things to check when a tunnel is Down.
It was started by hand and the terminal closed.
cloudflared tunnel run myappin an SSH session stops when the session ends. Install it as a service so it survives logouts and reboots.The service can't find its config. By default cloudflared looks for
~/.cloudflared/config.yml. Becausesudocan change$HOMEto/root, Cloudflare's Linux service docs suggest passing the path explicitly when installing:sudo cloudflared --config /home/you/.cloudflared/config.yml service installThe credentials file is missing. The config's
credentials-filemust point at the<UUID>.jsonfile created bycloudflared tunnel create. If you moved servers and didn't copy it, the tunnel can't authenticate.Outbound traffic is blocked. cloudflared connects out to Cloudflare on port 7844, over TCP (
http2) and UDP (quic). Per Cloudflare's firewall page, allow both toregion1.v2.argotunnel.comandregion2.v2.argotunnel.com. A quick check from the server:dig A region1.v2.argotunnel.com nc -vz region1.v2.argotunnel.com 7844
To see more while you debug, run it in the foreground with a log level, or stream logs from Cloudflare's side:
cloudflared tunnel --loglevel debug run myapp
cloudflared tail <UUID>
The dashboard also has a Live logs tab on the tunnel page, if your account role has edit permissions for tunnels.
Once the service starts and connects, the tunnel turns Healthy and the 1033 should clear within moments.
Step 3: tunnel is Healthy but you still get 1033
Cloudflare's guidance here is short: "confirm your DNS record points to the correct tunnel subdomain". In practice this happens because the DNS record and the tunnel are separate objects. The DNS routing docs say so directly: "The DNS record and the tunnel are independent." Deleting or recreating a tunnel doesn't update the CNAME for you.
How it goes wrong:
- You deleted a tunnel and created a new one with the same name. The name matches, but the new tunnel has a new UUID. The CNAME still points at the old
<old-UUID>.cfargotunnel.com. - The DNS record lives in a different Cloudflare account. The docs state that the
cfargotunnel.comsubdomain only proxies traffic for DNS records in the same Cloudflare account as the tunnel. - A typo or copy-paste error in a hand-made CNAME.
To check, open DNS > Records for your zone in the dashboard and read the CNAME target for the hostname. (A public dig lookup won't help, because a proxied record resolves to Cloudflare's IPs, not to the CNAME target.) Compare the UUID with the ID from cloudflared tunnel list. Deleted tunnels can be listed too:
cloudflared tunnel list -d
To fix a locally managed tunnel, point the hostname at the right tunnel. The --overwrite-dns flag (short form -f) replaces an existing record with the same name:
cloudflared tunnel route dns --overwrite-dns myapp app.example.com
For a tunnel you manage from the dashboard, edit the hostname route on the tunnel's page, or edit the CNAME target in DNS records by hand.
One related note: the same DNS docs say that when a tunnel stops, its DNS record remains and visitors may see error 1016 instead. If you see 1016 on a hostname you know is served by a tunnel, work through Step 1 and Step 2 the same way.
Quick checklist
- Dashboard Networking > Tunnels, or
cloudflared tunnel list: what's the status? - Inactive or Down:
systemctl status cloudflared, readjournalctl -u cloudflared, start it. - Check the config path, credentials file and outbound port 7844 (TCP and UDP).
- Healthy but 1033: compare the CNAME target's UUID with the tunnel's UUID, in the same account.
- Fix the record with
cloudflared tunnel route dns --overwrite-dnsor in the dashboard. - Seeing 502 instead? The tunnel is fine; fix the service URL or the app behind it.
If you're choosing between a tunnel and a private network for this app, Cloudflare Tunnel vs Tailscale compares the two.
Most 1033s come from a tunnel running on a machine that isn't always on. EasySpawn runs your app on a persistent server that already has a public URL, SSL and your own domain, so you don't need a tunnel to put it online. If you still want one (for Cloudflare Access, say), cloudflared runs there as a normal service that stays up when your laptop is closed. Join the waitlist.
Related: Cloudflare Tunnel Explained: Put a Local App Online Without Opening Ports · Cloudflare Tunnel vs Tailscale: Public Exposure or Private Network? · DNS Records Explained: A, CNAME, MX, and TXT for Beginners · What Is Cloudflare? What It Does When You Put Your Site Behind It
Keep reading
Cloudflare Error 520 "Web Server Is Returning an Unknown Error": Causes and Fixes
Cloudflare error 520 means your origin sent an empty, malformed or cut-off response. The common causes, how to bypass Cloudflare to test, and how to fix it.
ERR_SSL_PROTOCOL_ERROR: What It Means and How to Fix It
ERR_SSL_PROTOCOL_ERROR means the browser couldn't complete a secure HTTPS connection because the server's response wasn't valid TLS. The usual causes — HTTPS pointed at a port speaking plain HTTP, a missing certificate, localhost over https, Cloudflare SSL modes — and how to fix each.