Carter Kozak
@ckozak.net
Passionate about software infrastructure, observability, static analysis, and java performance.
https://github.com/carterkozak
https://github.com/carterkozak
I knew you made incredible things with Java, but this is another level!
I've been working on my latte for the past two years and can just barely get a consistent monks head. So much respect :-)
I've been working on my latte for the past two years and can just barely get a consistent monks head. So much respect :-)
November 23, 2024 at 3:28 PM
I knew you made incredible things with Java, but this is another level!
I've been working on my latte for the past two years and can just barely get a consistent monks head. So much respect :-)
I've been working on my latte for the past two years and can just barely get a consistent monks head. So much respect :-)
Yes, ckozak@apache.org
November 22, 2024 at 12:56 AM
Yes, ckozak@apache.org
Reposted by Carter Kozak
Rich pre-crash recordings, I would say. JFR has a reputation as performance tool, but I've heard the original thinking was also, "collect past data in circular buffer, dump it along with a bug report, let JVM support people look at what happened recently".
November 19, 2024 at 4:33 PM
Rich pre-crash recordings, I would say. JFR has a reputation as performance tool, but I've heard the original thinking was also, "collect past data in circular buffer, dump it along with a bug report, let JVM support people look at what happened recently".
This!
We rely heavily on JFR to read into crashes, oomkills, etc. A crash report (if one exists, not so lucky for oomkills!) gives us data from one distinct event, where the JFR paints a landscape to contextualize the causes.
We rely heavily on JFR to read into crashes, oomkills, etc. A crash report (if one exists, not so lucky for oomkills!) gives us data from one distinct event, where the JFR paints a landscape to contextualize the causes.
November 19, 2024 at 9:12 PM
This!
We rely heavily on JFR to read into crashes, oomkills, etc. A crash report (if one exists, not so lucky for oomkills!) gives us data from one distinct event, where the JFR paints a landscape to contextualize the causes.
We rely heavily on JFR to read into crashes, oomkills, etc. A crash report (if one exists, not so lucky for oomkills!) gives us data from one distinct event, where the JFR paints a landscape to contextualize the causes.