Главная
Study mode:
on
1
Intro
2
Target audience
3
Agenda
4
Scheduling & Memory Management
5
Memory Management: Stack
6
Memory Management: Heap
7
CPU Profiler: SIGPROF for every 10ms of CPU Time
8
CPU Profiler: Add Stacktrace to Profile
9
CPU Profiler: Waiting for Go 1.18
10
Block Profiler: Channels
11
Block Profiler: Mutexes
12
Mutex Profiler
13
Block vs Mutex
14
Quick Note on Time: 1 Request - 3 Goroutines
15
Memory Profiling
16
Goroutine Profiling
17
Tracing Manually
18
Runtime Tracer
19
Profiling and Tracing: A Faustian Bargain?
20
Profiling and Tracing Overhead Analysis
21
Overhead Analysis: SQL Workload
22
Overhead Analysis: HTTP Workload
23
Overhead Analysis: JSON Workload
24
JsoN Workload (Concurrency = 8)
25
Overhead Analysis: Chan Workload
26
Overhead Analysis: Error Sources
27
Metrics
28
Linux perf
29
bpftrace
30
Scheduling / Execution Observability
31
Memory Management Observability
32
Recap
Description:
Explore profiling and observability techniques for Go applications in this comprehensive conference talk from GopherCon 2021. Learn how to reduce costs, improve latency, and debug issues like memory leaks and performance regressions. Gain insights into the Go runtime's unique features and understand the built-in CPU, Memory, Block, Mutex, and Goroutine Profilers, as well as the Runtime Tracer. Discover advanced tools such as eBPF, Linux Perf, and Delve for optimizing and debugging Go applications. Follow along as the speaker covers topics including scheduling, memory management, profiling techniques, tracing, overhead analysis, and observability tools, providing a solid foundation for improving Go application performance and troubleshooting skills.

Go Profiling and Observability from Scratch

Gopher Academy
Add to list
0:00 / 0:00