Brendan Gregg

Brendan Gregg
Name	Brendan Gregg
Occupation	Computer engineer, performance analyst, author
Known for	Systems performance, DTrace, performance visualization
Awards	ACM SIGOPS Hall of Fame, USENIX LISA Top Contributors

Brendan Gregg Brendan Gregg is a computer engineer and performance analyst known for contributions to systems performance, performance engineering tooling, and performance visualization. He has worked on performance in large-scale environments at institutions such as Sun Microsystems, Oracle Corporation, Netflix, and IBM. Gregg authors technical books and maintains extensive public collections of performance patterns, methodologies, and open-source tools used across the Unix and Linux ecosystems.

Early life and education

Gregg was born in Australia and completed tertiary studies in computing and engineering. He pursued studies at institutions including University of New South Wales and engaged with academic communities that intersect with research groups in operating systems, computer architecture, and programming languages. Early influences included seminal works from researchers associated with BSD, Solaris, and the broader Unix tradition.

Career

Gregg’s industry career began with roles at Sun Microsystems where he contributed to performance engineering efforts related to Solaris and the DTrace project. After Sun Microsystems merged with Oracle Corporation, he continued work on tracing and observability. He later joined Joyent and subsequently Netflix, where he focused on performance for large-scale cloud services and distributed systems. Gregg has also worked with IBM and provided consultancy and talks at conferences such as USENIX, ACM SIGOPS, and ISPASS.

Throughout his career Gregg has engaged with standards and communities including OpenTracing, Linux Foundation, and cloud providers like Amazon Web Services, Google Cloud Platform, and Microsoft Azure through talks, tutorials, and open-source contributions. He has collaborated with engineers from projects such as ZFS, Kubernetes, Docker, and eBPF implementations.

Research and contributions

Gregg’s research focuses on performance analysis, observability, low-overhead tracing, and systems-level performance visualization. He contributed to the DTrace dynamic tracing framework developed for Solaris and worked on applying tracing techniques to Linux and other platforms. Gregg popularized methodologies for performance investigation such as flame graphs, latency histograms, and system-level performance patterns. His work intersects with kernel subsystems like Linux scheduler, ksh, and ZFS, and with tools including perf, SystemTap, ftrace, and eBPF.

Gregg developed and evangelized performance visualization techniques—most notably flame graphs—that help diagnose CPU and latency issues in complex stacks involving Java, Go, Node.js, C++, and Python applications. He analyzed performance phenomena arising in environments comprised of SSD storage, NVMe, virtualized platforms like VMware, and container orchestration with Kubernetes.

His pattern catalogs and methodology documents connect observed symptoms to root causes across networking stacks involving TCP/IP, UDP, and BPF filters, as well as I/O subsystems using POSIX interfaces. Gregg’s investigations have informed operational practices for companies running at web scale and have influenced research published at venues like USENIX Annual Technical Conference and ACM SIGMETRICS.

Publications and tools

Gregg is author of multiple books and numerous technical papers, tutorials, and blog posts. Key works include "Systems Performance" and "BPF Performance Tools", which synthesize profiling techniques, kernel internals, and monitoring strategies. He maintains open-source tooling and examples such as flame graph generators, BPF-based tracing utilities, and collection scripts that integrate with Prometheus, Grafana, and other observability stacks. Gregg’s tools interact with kernel interfaces including eBPF, perf_event_open, and tracing frameworks like DTrace and SystemTap.

He has presented at conferences such as USENIX LISA, Velocity Conference, KubeCon, QCon, and Black Hat and contributed chapters and tutorials to collections produced by O’Reilly Media and academic proceedings from ACM and IEEE. His published repositories provide reproducible examples for profiling applications in languages and runtimes like Java Virtual Machine, Linux Containers, and FreeBSD environments.

Awards and recognition

Gregg’s work has received recognition from professional organizations and conferences. He has been cited in awards such as ACM-related honors and has been included in community acknowledgments like USENIX LISA contributor lists and ACM SIGOPS Hall of Fame citations. His flame graph technique and associated utilities have become standard references in operational incident postmortems across companies including Netflix, Google, Facebook, and cloud providers. Gregg’s books and tutorials are widely used in university courses on operating systems and performance engineering at institutions like Massachusetts Institute of Technology, Stanford University, and University of California, Berkeley.

Personal life

Gregg lives in United States while maintaining ties to Australia and participates in open-source communities and technical mentoring programs. He contributes to public documentation, talks at meetups associated with groups like Linux Foundation chapters, and collaborates with engineers from projects such as FreeBSD, OpenBSD, and NetBSD. Outside technical work, he engages with community education efforts and supports initiatives that improve software reliability and operational practices.

Category:Computer engineers Category:Systems performance