GC duration and pause time grow linearly

[Aleksey Shipilev]

On 6/1/20 5:51 AM, Kirill Ilyukhin wrote:
> Please look at the attached images, these are charts built by cgeasy.io for
> my GC log. GC/pause time grows as if it were a linear function of JVM
> uptime. Is this an expected behavior? I am not complaining, performance is
> good enough, but GC/pause time trend looks a bit suspicious.

If application does not change its load profile, then it is not the intended behavior.

> This JVM is a standby instance of Tomcat 9.0.22 as a Windows service, doing
> almost nothing. Most of GC are time-triggered, every 5 minutes.
> Please let me know if you need more information.

Unfortunately, JDK 13 is past its support lifetime. And it lacks the more extensive per-cycle
logging to really follow up on this. Take either 14u or 11u nightlies?

> [image: image.png]
> [image: image.png]
> [image: image.png]

Attachments are stripped on this list, please post URLs to somewhere else? Verbatim GC logs would be
even better, from either 14u or 11u nightlies.

-- 
Thanks,
-Aleksey