[Bug]: [Sentry] ServerWebInputException POST /api/v1/actions/execute | APPSMITH-SERVER-E

[rohan-arthur]

Is there an existing issue for this?

• I have searched the existing issues

Description

https://appsmith.sentry.io/issues/4303877504

Steps To Reproduce

NA

Public Sample App

No response

Environment

Production

Issue video log

No response

Version

Cloud

[NilanshBansal]

Log with full error trace: https://appsmith.sentry.io/issues/4303877504/events/0dbba7673b9540d5b238d89f6a6b2a4d/?project=4505267188465664

[sumitsum]

Update:
I was able to reproduce it artificially by removing the body section from the HTTP request sent to the execute action API. However, I have not been able to figure out why something like this could occur in production environment. Based on the data available on Sentry I tried the following to reproduce the issue but couldn't:

1. Supply large payload evaluated value for dynamic binding when executing query.
2. Change browser version to older Chrome v119.
3. Try to run / edit application via Andorid mobile phone.
4. Try to run some queries after importing template application or generate CRUD app from DB.

I also found that a bug (different but similar) was fixed with Spring framework multipart code and some more fixes have gone in recently. Maybe upgrading the spring boot version could help here.

[NilanshBansal]

@sumitsum does logging of this bug to sentry makes sense? Should we keep this as a log only in mezmo and remove the logging of this to sentry to preserve our sentry quota?
cc: @rohan-arthur

[sumitsum]

Hi @NilanshBansal it could still be useful to keep sending it to Sentry. The stack trace for this issue gets printed on Mezmo even now, however, AFAIK, there is no known mechanism in Mezmo to provide us alert or stats for the occurrence of this error. In case the occurrence of this error increases sharply in future we may want that to be visible on Sentry. If it is only on Mezmo, then the issue may get ignore even when the severity increases.