You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
If an application contains an instrumented method that has non-ascii characters (in the assembly name, namespace, type name, or method name) it can cause an application running on Linux to crash.
The problem appears to be related to the way we log things within the profiler. On Linux, if we attempt to log a message that contains a non-ascii character it causes the app to crash. On Windows, it partially writes the log message (up until the non-ascii character) and then everything continues.
Current workarounds include:
Changing the log level so that the profiler doesn't attempt to log non-ascii characters
Removing the non-ascii characters from the code
Expected Behavior
The agent should not crash and the method should be instrumented normally. The log messages should support non-ascii characters.
The issue doesn't appear if the log level is reduced to not include the problematic log lines.
The issue doesn't appear on Windows.
Steps to Reproduce
Create a new aspnetcore 7 application that is dockerized and runs in Linux container.
Add a controller endpoint that contains a non-ascii character (such as \u0435)
Enable finest level logs for the agent (to ensure that the profiler attempts to log the non-ascii character).
Send a request to the endpoint with the non-ascii character
Notice that the docker container terminates with a non-successful exit code (which indicates that the app crashed)
Your Environment
This was reproduced in a Linux docker container but it can probably be reproduced on any Linux system. This can probably also be reproduced when using the Transaction or Trace attributes (or xml instrumentation) on a method containing non-ascii characters.
For Maintainers Only or Hero Triaging this bug Suggested Priority (P1,P2,P3,P4): P4 Suggested T-Shirt size (S, M, L, XL, Unknown): Unknown
The text was updated successfully, but these errors were encountered:
Description
If an application contains an instrumented method that has non-ascii characters (in the assembly name, namespace, type name, or method name) it can cause an application running on Linux to crash.
The problem appears to be related to the way we log things within the profiler. On Linux, if we attempt to log a message that contains a non-ascii character it causes the app to crash. On Windows, it partially writes the log message (up until the non-ascii character) and then everything continues.
Current workarounds include:
Expected Behavior
The agent should not crash and the method should be instrumented normally. The log messages should support non-ascii characters.
Troubleshooting or NR Diag results
The issue doesn't appear if the log level is reduced to not include the problematic log lines.
The issue doesn't appear on Windows.
Steps to Reproduce
\u0435
)Your Environment
This was reproduced in a Linux docker container but it can probably be reproduced on any Linux system. This can probably also be reproduced when using the Transaction or Trace attributes (or xml instrumentation) on a method containing non-ascii characters.
For Maintainers Only or Hero Triaging this bug
Suggested Priority (P1,P2,P3,P4): P4
Suggested T-Shirt size (S, M, L, XL, Unknown): Unknown
The text was updated successfully, but these errors were encountered: