Set trace level to info

[betzw]

Description

@SeppoTakalo should improve situation for issue #247.

[betzw]

Can someone pls. let me know why several of the checks are failing (seems as if I have no access to the Details)?

[SeppoTakalo]

Due the security reasons, some of our CI jobs don't execute tests from pull requests coming outside of these repositories. Therefore they fail.

Somebody from Client team should review this and replicate the PR into a local branch. @yogpan01 @teetak01 can you help on this?

[teetak01]

Well, the PR itself is fine, but is this really the right way to "fix" this? There seems to an issue that the mesh cannot handle the activity. This is only hiding it.

[betzw]

Don't think that there is a way to fix this issue as it basically is a timing issue which is caused by the huge amount of debug messages when using trace level debug.
On the other hand, setting the trace level to info would allow users to be able to see what's going on in the stack while still having working connections.

[teetak01]

Could same kind of delays be caused if the application is doing some heavy processing (like TLS operations) or some heavy customer application decoding?

[betzw]

In theory yes, but it must be very heavy processing ... I mean producing debug output over printf and the like is so time consuming because it typically is implemented in polling mode, i.e. busy waiting for each character to be sent over the serial line, inhibiting any other kind of processing to proceed.

[JanneKiiskila]

Is there a conclusion on this? @yogpan01 @SeppoTakalo

[betzw]

As far as I know, not really.
What I can add is that this would have been very helpful while we worked on #266.

[teetak01]

Part of #327 already.