This is running in Tomcat 8.0.33 in a Docker Container via AWS’s ECS which is
identical to before with log4j2 2.5.
We’ve run the application with Visual VM and the one thing for sure that we’ve
seen is that in 2.6.2 it created tons of threads, something like 50+ but on 2.6
or 2.7 it was only 2. Now the threads were short lived, but they got created.
We’re running another performance test today with Async loggers to see if that
helps or exhibits the same thing though previous testing with Async had some of
the same char
Here’s our layout pattern:
[%t] %d{DATE} %-5p %-15c{1} [%X]: %cm%n
Please not the %cm is a custom message handler that we use to use to handle
security filtering of the message payload aka we extend from
LogEventPatternConverter.
Thanks,
Ron DiFrango
On 2/13/17, 11:22 AM, "Matt Sicker" <[email protected]> wrote:
What server environment are you running this in?
On 13 February 2017 at 09:19, Remko Popma <[email protected]> wrote:
> Ron,
>
> We haven't heard of any issues like you describe.
> Have you tried running your application with Java Flight Recorder
> <https://docs.oracle.com/javacomponents/jmc-5-4/jfr-
> runtime-guide/run.htm#JFRUH176>?
> This should help diagnose what is going on.
>
> Remko
>
>
>
> On Mon, Feb 13, 2017 at 9:59 PM, DiFrango, Ronald <
> [email protected]> wrote:
>
> > All,
> >
> > We recently upgrade to 2.6 and noticed a dramatic increase in CPU and
> > Thread utilization that seems to be tied to the new “garbage free” mode
> of
> > log4j 2.6. Here’s some of the baseline numbers:
> >
> >
> > · Log4j 2.5: CPU typically ran around 25%
> >
> > · Log4j 2.6: CPU typically ran around 75%
> >
> > · Log4j 2.6.2+: CPU typically ran around 100%
> >
> > · We’ve also tried turning off garbage free mode and that made
> > things worse as the CPU was around 120% and caused us to not meet our
> SLA’s
> >
> > It important to note that this is a REST Api using Jersey and typically
> > responds in about in under 50ms on a per request so its high volume, but
> > the logging level is WARN or higher except for our single performance
log
> > record which is written once per request using the lambda base approach.
> >
> > Our next test is going to be switching to all ASYNC loggers and see what
> > effect that has, but I guess the general question is, has anyone else
> seen
> > this? Any thoughts?
> >
> > Ron DiFrango
> > ________________________________________________________
> >
> > The information contained in this e-mail is confidential and/or
> > proprietary to Capital One and/or its affiliates and may only be used
> > solely in performance of work or services for Capital One. The
> information
> > transmitted herewith is intended only for use by the individual or
entity
> > to which it is addressed. If the reader of this message is not the
> intended
> > recipient, you are hereby notified that any review, retransmission,
> > dissemination, distribution, copying or other use of, or taking of any
> > action in reliance upon this information is strictly prohibited. If you
> > have received this communication in error, please contact the sender and
> > delete the material from your computer.
> >
>
--
Matt Sicker <[email protected]>
________________________________________________________
The information contained in this e-mail is confidential and/or proprietary to
Capital One and/or its affiliates and may only be used solely in performance of
work or services for Capital One. The information transmitted herewith is
intended only for use by the individual or entity to which it is addressed. If
the reader of this message is not the intended recipient, you are hereby
notified that any review, retransmission, dissemination, distribution, copying
or other use of, or taking of any action in reliance upon this information is
strictly prohibited. If you have received this communication in error, please
contact the sender and delete the material from your computer.
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
