https://prefect.io logo
Join Slack
Powered by
<@ULVA73B9P> I am using `Prefect 3.2.6` and have ...
# ask-marvin
t
@Marvin I am using 
Prefect 3.2.6
and have currently problem deploying flow on ther server with error below
Copy code
File "/home/vua/miniconda3/envs/cenv/lib/python3.11/site-packages/httpcore/_sync/connection.py", line 101, in handle_request
    raise exc
  File "/home/vua/miniconda3/envs/cenv/lib/python3.11/site-packages/httpcore/_sync/connection.py", line 78, in handle_request
    stream = self._connect(request)
             ^^^^^^^^^^^^^^^^^^^^^^
  File "/home/vua/miniconda3/envs/cenv/lib/python3.11/site-packages/httpcore/_sync/connection.py", line 124, in _connect
    stream = self._network_backend.connect_tcp(**kwargs)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/vua/miniconda3/envs/cenv/lib/python3.11/site-packages/httpcore/_backends/sync.py", line 208, in connect_tcp
    sock = socket.create_connection(
           ^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/vua/miniconda3/envs/cenv/lib/python3.11/socket.py", line 848, in create_connection
    sock.connect(sa)
What could I do ?
m
✅ thought for 258.7 seconds
Thanks for sharing the traceback. That stack trace is from httpcore failing to open a TCP socket, which almost always means the Prefect client/CLI can’t reach your Prefect API (wrong URL, server not running, port closed, DNS/proxy issues, etc.). A quick checklist to isolate the issue: 1) Confirm which API you’re targeting - Are you using Prefect Cloud or a self-hosted Prefect Server? - What is your current 
PREFECT_API_URL
? Run:
Copy code
prefect config view --show-sources
Look for 
PREFECT_API_URL
and (if using Cloud) that you have a 
PREFECT_API_KEY
. 2) Test the health endpoint directly - In Prefect 3.x the health endpoint is 
<PREFECT_API_URL>/health
. - Example (self-hosted default): if 
PREFECT_API_URL=http://<server-host>:4200/api
, test:
Copy code
curl -v http://<server-host>:4200/api/health
- Docs: Health check endpoint If this does not return healthy (HTTP 200 / JSON), the issue is network/server-side rather than your deployment code. 3) If you are self-hosting Prefect Server - Make sure the server is running and listening on a reachable interface:
Copy code
prefect server start --host 0.0.0.0 --port 4200 --ui
- Ensure your client can reach that host/port (firewall/security group allows 4200, correct DNS/IP). - Set your client to use that API URL:
Copy code
prefect config set PREFECT_API_URL="http://<server-host>:4200/api"
- From the client, re-test:
Copy code
curl -v $PREFECT_API_URL/health
4) If you are using Prefect Cloud - Use the CLI to configure your profile (this sets 
PREFECT_API_URL
and uses your key):
Copy code
prefect cloud login -k <YOUR_API_KEY> -w <account_handle>/<workspace_handle>
- Re-test:
Copy code
curl -v $(prefect config view | grep PREFECT_API_URL | awk '{print $3}')/health
- Cloud connection guide: Connect to Cloud 5) Common environment gotchas - Proxies: if behind a corporate proxy, ensure 
NO_PROXY
includes your Prefect host (e.g., 
NO_PROXY=localhost,127.0.0.1,<server-host>
). Otherwise, TCP may time out or be blocked. - HTTPS/self-signed certs: if using HTTPS on your own server, ensure the certificate is trusted by the client OS or use a trusted reverse proxy in front of Prefect. If the above checks don’t resolve it, could you share: - Whether you’re targeting Cloud or self-hosted Server - The exact value of 
PREFECT_API_URL
(you can redact host if needed) - The exact command/code you’re using to deploy (e.g., 
prefect deploy
or 
flow.from_source(...).deploy(...)
) - The complete error line after 
sock.connect(sa)
(e.g., “Connection refused”, “Timed out”, “Name or service not known”, etc.) With that info I can pinpoint the fix quickly.
Open in Slack
PreviousNext
Powered by