Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
eBPF in Microservices Observability
Search
JBD
August 18, 2021
Programming
1
1.7k
eBPF in Microservices Observability
JBD
August 18, 2021
Tweet
Share
More Decks by JBD
See All by JBD
eBPF in Microservices Observability at eBPF Day
rakyll
1
2.1k
OpenTelemetry at AWS
rakyll
1
1.8k
Debugging Code Generation in Go
rakyll
5
1.5k
Are you ready for production?
rakyll
8
2.7k
Servers are doomed to fail
rakyll
3
1.5k
Serverless Containers
rakyll
1
240
Critical Path Analysis
rakyll
0
550
Monitoring and Debugging Containers
rakyll
2
1.1k
CPDD
rakyll
0
4.1k
Other Decks in Programming
See All in Programming
CSS Linter による Baseline サポートの仕組み
ryo_manba
1
100
時計仕掛けのCompose
mkeeda
1
300
Amazon ECS とマイクロサービスから考えるシステム構成
hiyanger
2
560
Ruby on cygwin 2025-02
fd0
0
140
ペアーズでの、Langfuseを中心とした評価ドリブンなリリースサイクルのご紹介
fukubaka0825
2
320
Pythonでもちょっとリッチな見た目のアプリを設計してみる
ueponx
1
560
How mixi2 Uses TiDB for SNS Scalability and Performance
kanmo
37
14k
Grafana Loki によるサーバログのコスト削減
mot_techtalk
1
130
2,500万ユーザーを支えるSREチームの6年間のスクラムのカイゼン
honmarkhunt
6
5.3k
Rubyで始める関数型ドメインモデリング
shogo_tksk
0
110
WebDriver BiDiとは何なのか
yotahada3
1
140
プログラミング言語学習のススメ / why-do-i-learn-programming-language
yashi8484
0
130
Featured
See All Featured
How to Think Like a Performance Engineer
csswizardry
22
1.3k
Responsive Adventures: Dirty Tricks From The Dark Corners of Front-End
smashingmag
251
21k
Reflections from 52 weeks, 52 projects
jeffersonlam
348
20k
Save Time (by Creating Custom Rails Generators)
garrettdimon
PRO
29
1k
[RailsConf 2023] Rails as a piece of cake
palkan
53
5.2k
Design and Strategy: How to Deal with People Who Don’t "Get" Design
morganepeng
129
19k
Visualization
eitanlees
146
15k
The Cost Of JavaScript in 2023
addyosmani
47
7.3k
What’s in a name? Adding method to the madness
productmarketing
PRO
22
3.3k
A Philosophy of Restraint
colly
203
16k
Visualizing Your Data: Incorporating Mongo into Loggly Infrastructure
mongodb
45
9.4k
Building an army of robots
kneath
303
45k
Transcript
@rakyll eBPF in Microservices Observability Jaana Dogan Principal Engineer, AWS
[email protected]
@rakyll About me • Not a Linux developer. • Working
on monitoring, observability and performance. • Multi-tenancy and microservices focus.
@rakyll
@rakyll How does eBPF work? process JIT compiler Verifier Sockets
TCP/IP BPF Maps code (accessible from the user space)
@rakyll Where can eBPF hook into? - Kernel and user
functions - System calls - Network events - Kernel tracepoints
@rakyll Challenges in microservices
@rakyll Challenges in microservices We don’t just monitor VMs or
processes. We monitor critical paths.
@rakyll What’s next? service service database storage service
@rakyll What’s next? service service database storage service
@rakyll Challenges in microservices Context matters. Downstream stack don’t have
context.
@rakyll What’s next? process Linux kernel process process M:N Problem
@rakyll What’s next? process Linux kernel process process RPCs M:N
Problem
@rakyll What’s next? process Linux kernel process process RPCs container
container M:N Problem
@rakyll What’s next? process Linux kernel process process RPCs container
container Kubernetes pod, ECS task M:N Problem
@rakyll Challenges in microservices We initially debug RPCs. We debug
functions or syscalls secondarily.
@rakyll Challenges in microservices Too much data. Need runtime controls
to modify the collection.
@rakyll Challenges in microservices Instrumentation is a two-year roadmap. Data
is not consistent.
@rakyll Networking observability is core. Out of the box instrumentation
is essential. Extensibility in runtime is critical. Decoration and enrichment is needed.
@rakyll How does eBPF help?
@rakyll Network Diagnostics TCP, UDP, HTTP, gRPC metrics Inspect protocols
(MySQL, Postgres, ...)
@rakyll Service Maps
@rakyll Distributed Traces Automatically create request span if a trace
header is present. GET /users HTTP/1.1 Host: users.service Accept-Encoding: gzip, deflate Connection: Keep-Alive Traceparent: 00-4bf92f3577b34da6a3ce929d0e0e4736-00f067aa0ba902b7-01
@rakyll Fleet-wide Profiling context kernel
@rakyll Decorating with Context eBPF agent process JIT compiler Verifier
Sockets TCP/IP BPF Maps API Server
@rakyll Runtime Extensibility eBPF agent process JIT compiler Verifier Sockets
TCP/IP BPF Maps code
@rakyll Examples - Cillium/Hubble - Pixie - Flowmill
@rakyll What’s next? - High level language to write probes.
- Make eBPF agents widely available. - More platforms supporting eBPF. - Reusable eBPF event processing.
@rakyll Thank you Jaana Dogan
[email protected]