eBPF / Linux Performance Specialist
eBPF and Linux Performance Specialists instrument the kernel and userspace to diagnose latency, throughput, and resource-contention issues at the deepest level of the stack. Their English covers writing performance analysis reports, presenting flamegraph findings to engineering leadership, and documenting custom eBPF-based observability tooling. This path builds the specialised vocabulary to communicate kernel tracing work clearly to mixed audiences.
Topics covered
- eBPF Programming Model
- Kernel Tracing with kprobes & tracepoints
- Flamegraph Analysis
- perf & BCC / bpftrace Tooling
- Performance Analysis Reports
- Linux Scheduler & Memory Vocabulary
Vocabulary spotlight
4 terms every eBPF / Linux Performance Specialist should know in English:
Extended Berkeley Packet Filter — a kernel technology that allows sandboxed programmes to run in the Linux kernel without changing kernel source code or loading kernel modules
"We use an eBPF programme attached to the TCP send path to measure per-connection latency without any code changes."
A visualisation of hierarchical profiling data in which the width of each frame represents the proportion of total time spent in that code path
"The flamegraph revealed that 40% of CPU time was consumed by lock contention in the memory allocator."
A dynamic tracing mechanism in the Linux kernel that allows a handler function to be inserted at almost any kernel instruction address
"A kprobe on the block I/O completion handler lets us measure individual disk request latency without recompiling the kernel."
Describing time a thread spends waiting — blocked on I/O, locks, or scheduling — rather than executing on a CPU core
"Off-CPU analysis showed that database threads were spending 70% of their time waiting for storage I/O."
📚 Vocabulary Reference
Key terms organised by category for eBPF / Linux Performance Specialists:
eBPF & Tracing
Profiling & Visualisation
Linux Kernel Concepts
Performance Metrics & Reporting
Recommended exercises
Real-world scenarios you'll practise
- Writing a performance analysis report that presents flamegraph findings to engineering leadership.
- Explaining an eBPF-based observability tool to a site-reliability team unfamiliar with kernel tracing.
- Documenting a kprobe instrumentation approach in a runbook for on-call engineers.
- Presenting CPU-scheduling and off-CPU analysis results during a production incident post-mortem.