High high memory usage traced to ScopedLogger

[solomon-okolie]

APM Agent version

1.25.1

Environment

Operating system and version: WindowsServer, Windows10

.NET Framework/Core name and version : .NET 6

Application Target Framework(s): .NET 6

Describe the bug

We have console a app that processes messages from a queue and we capture transactions manually like below
Agent.Tracer.CaptureTransaction("TestProcessor", "MQP", () => { // code here }); and it's eating quite a lot of memory after a couple of hours. This looks very similar to #2145. but this time happening in ScopedLogger.

We did a little test based on the commit fix for #2145..

Before change

Heap snapshot after 11,000 transactions, and a forced a GC

After change

Heap snapshot after 11,000 transactions, and a forced a GC

[Mpdreamz]

Thanks for raising this, I added an initial fix for this here:

#2242

Including a few other logging related fixes.

I'll do a pass over our logging to find any high cardinality suspects too.

[solomon-okolie]

Any idea when a new nuget version will be available?

[AlfonsRoerdink]

I will be great if a new hotfix release can be made because my pods are crashing after a few days because out-of-memory.
Is there an estimation when the new nuget version will be available?

[Mpdreamz]

This is now released under 1.25.3.

So sorry and apologies you got bit too by this @AlfonsRoerdink let us know if the update did not resolve your issues.

[AlfonsRoerdink]

I installed the new version and the application is already running 16 hours without a remarkable memory leakage,
so it seems to be fixed! Thank you very much!

[Mpdreamz]

Thanks a ton for looping back with us!

I will go into the weekend much more relaxed now :)