Hi being using HTTP4k in production for roughly a ...

# http4k

[krtko]

Hi being using HTTP4k in production for roughly a year now and its been great! So thank you.

I'm having a very odd and very rare issue.
I recently updated from 3.183.0 to 3.249.0 and I use Undertow in production
About once a day, a single request throw this error io.undertow.request - UT005007: Request was not fully consumed
Any ideas?

[dave]

That's very strange. I've had a quick look at the history of the Undertow ServerConfig code and there haven't been any (relevant) changes to it apart from version bumps of the Undertow libraries from "2.0.28.Final" to "2.1.3.Final". So TBH, it looks like something that is due to Undertow.

If you'd like to prove that, you could force a downgrade in the Undertow JAR version to the previous one and see if the problem disappears. If it does then maybe raise an issue with Undertow?

[krtko]

Thanks I will try that. What backend do you typically use in production?

[dave]

we always use Undertow as a first choice. unless we need Websockets in which case Jetty is the only option currently.

[krtko]

Ok thank you

[krtko]

I will try down grading the server when I get an opportunity.

[dave]

👍

[dave]

please let us know if you find anything 🙂

[krtko]

Will do

[krtko]

Oddly enough since I posted this I haven't had one of these errors.
I also haven't changed anything since the errors began and they've been very regular.
I'm guessing at this point it was an issue with Heroku's router and their JVM package.
Thanks for the help and wonderful HTTP framework