That’s correct. anvil.server.call on all of them
This is another strange issue. Might just be on my side, but when I launch a server console from the editor and try to run import anvil.server manually, it loads forever.
The last test went for 30 seconds before I terminated the console. Perhaps the origin of this issue is hidden somewhere in this on the server side?
— disregard. the application just took a long time to load.
Thanks for this too! Your previous info has really helped, and we’re making progress towards understanding the problem now. I will post another update soon.
- Python 3.10Beta (Standart) + simple “anvil.server.call” with request to datatable- Delays and Timeout errors
- Python 3.10Beta (Standart) + simple “anvil.server.call” - Delays 10-30 sec
- Full python 3 with the same test APP - works OK
APP ID: XMLQGQ5TR5DJWPBL
Session ID: W2WTFCR7CIMIMQY3TSRHHPFMZG7EVXYZ
22/01/2024 19:23 EET
Hi everyone.
We have now deployed a workaround that should speed things up in Python 3.10. We’ll continue investigating to identify and fix the root cause, but in the meantime things should be quite a lot faster now. I’ll continue to post updates to this thread. Thanks for your patience!
2 Likes
This may be a little late… however…
Profession Plan
Python 3.10 (Standard)
app id: C6BEDUQBLFLL4OU7
session id: PZWJOZN3KSVZHN3K55WKZG6I76K7HLYU
log error: user-CRITICAL 2024-01-22 09:03:17: Server code took too long
This error occurrs on the app’s first server access.
The app has been in production since Jan 1, 2024 with a few session timeouts but nothing to worry about. This morning I got several critical errors (the app emails me when such errors occur). I tried the production version twice with the same error as above. Then I tried the development branch and it worked, although with slow response. Then when I tried the production branch it worked again, although it was sluggish.
The workaround appears to work for me. Thank you
Hi Ian, if it helps, I set up an hourly test to check the server status for my app. It runs a couple of functions and times each step to see if / when things fall over. These are server calls that run w/o issues normally, so they should only time out when there is something wrong.
I can share the results table if that helps ID specific times of timeouts plus what the app was trying to do. It goes back to Sunday eve (GMT).
(The dashboard is also at the bottom of the app sign in page here, app.dcdr.io if that’s easier).
The dashboard is looking good today after a ~3 hour series of errors yesterday.
app ID Y26N6JTYNQLGMR7Z
Andrew
PS - Sorry I dropped off the thread earlier
1 Like