François Joubaud
Dev leader at Worldline.
Passion for Java observability.
Strong experience in the payment industry (Reuters, Trema, Wallstreet Systems, ION Trading and Worldline) on critical application servers.
Author of Jeyzer: incident analysis solution for Java, open source project.
Jeyzer 3.2 is virtual thread focused (recording and analysis).

LOCATION: Bern
KEYWORDS: Language, Performance, Tools

SPEAKER: François Joubaud   COMPANY: Worldline
SLIDES: 240618_virtual-threads-monitoring.pdf

By pushing the runtime to its extreme limits, virtual threads requires some adapted observability: What are the available means as of today?

What should we measure and how? Let's review the state of the art on Java 21+ and cover some cases illustrated with analysis reports (generated with Jeyzer).

Based on those points, we will discuss a pragmatic adoption of virtual threads in projects, considering a path to production.

Links:

• Jeyzer report on which we went through to illustrate the different VT cases.
• Jeyzer report which shows a VT leakage (Unleash the beast scenario without any limitation)
• The virtual monitoring guide
• The Jeyzer demo code (used to illustrate the cases)
• This test permits to replicate the leakage issue on Windows.
• The VT recording which permits to generate the above VT leakage report, using the Jeyzer online analyzer or the Jeyzer analyzer on Docker

LANGUAGE: Talk: en / Slides: en

François Joubaud
Dev leader at Worldline.
Passion for Java observability.
Strong experience in the payment industry (Reuters, Trema, Wallstreet Systems, ION Trading and Worldline) on critical application servers.
Author of Jeyzer: incident analysis solution for Java, open source project.
Jeyzer 3.2 is virtual thread focused (recording and analysis).

---