Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Sign up for free
Menu
Search
Features
All features
Private URLs
Password Protection
Custom URLS
Scheduled publishing
Remove Branding
Restrict embedding
Deck Collections
Notes
Features
All features
Private URLs
Password Protection
Custom URLS
Scheduled publishing
Remove Branding
Restrict embedding
Deck Collections
Notes
Explore
Featured decks
Featured speakers
Programming
Technology
Storyboards
Explore
Featured decks
Featured speakers
Programming
Technology
Storyboards
Pricing
Search
Sign in
Sign up for free
Serverless for mere mortals
Search
Sponsored
·
SiteGround - Reliable hosting with speed, security, and support you can count on.
→
Nikolay Sverchkov
April 22, 2019
Programming
1.1k
1
Share
Embed
Copy iframe code
Copy JS code
Copy link
Start on current slide
Serverless for mere mortals
Nikolay Sverchkov
April 22, 2019
Transcript
119 slides
SERVERLESS FOR MERE MORTALS NIKOLAY SVERCHKOV @ssnickolay
None
evilmartians.com
None
SERVERLESS FOR MERE MORTALS
OR
HOW I REDUCED COSTS BY 250 TIMES USING SERVERLESS
PAST
IN-HOUSE SERVER A LONG TIME AGO
A LONG TIME AGO MAINTAINABILITY IN-HOUSE SERVER
DEDICATED SERVER NOT SO LONG AGO ovh.com
DEDICATED SERVER NOT SO LONG AGO SCALABILITY
CLOUDS NOT SO LONG AGO
CLOUDS NOT SO LONG AGO
CLOUDS NOT SO LONG AGO
CLOUDS NOT SO LONG AGO ? CPU ? RAM ?
DISK
http://bit.ly/2X9qnp5
None
?
KANYE WEST
None
~66.500 RUB
LOAD 0 25 50 75 100 0 1 2 3
4 5 6 7
LOAD 0 25 50 75 100 0 1 2 3
4 5 6 7
CAPACITY PLANNING
None
CLOUDS NOT SO LONG AGO CAPACITY PLANNING
MORE ABSTRACTION! move away from server assessment models
None
2014
HOW IT WORKS
; ; ; AWS Lambda 1. SLEEP SOURCE_CODE.ZIP S3
AWS Lambda 1. SLEEP 2. EVENT
AWS Lambda 1. SLEEP 2. EVENT 3. HANDLE
function(event, context) { console.log(event.key1); return “Success”; }
AWS Lambda 1. SLEEP 2. EVENT 3. HANDLE 4. RESPONSE
; ; ; AWS Lambda 1. SLEEP 2. EVENT 3.
HANDLE 4. RESPONSE 5. -> 1
HTTP ❤ EVENT
BLOG + STATIC PAGES RAILS 3; DO; NGINX ASSETS PIPELINE
User Lambda Markdown -> HTML
User S3 Lambda Assets Markdown -> HTML CloudFront
User S3 Lambda Assets Markdown -> HTML + CloudFront
NOTE 1: THE RIGHT SOLUTION (JAMstack) +
NOTE 2: CAUTION OPEN FLAME
NOTE 2: CAUTION OPEN FLAME
CAUTION OPEN FLAME
⏱ DEC 2018 % +
DEC 2018 MARCH 2019 % +
VS http://bit.ly/2Dd7R7X
SERVERLESS
SERVERLESS $ serverless create !--template aws-go !!==> serverless.yml
SERVERLESS $ serverless deploy $ serverless rollback !--timestamp timestamp $
serverless invoke -f functionName $ serverless logs -f functionName $ serverless create !--template aws-go !!==> serverless.yml
SERVERLESS $ serverless deploy $ serverless rollback !--timestamp timestamp $
serverless invoke -f functionName $ serverless logs -f functionName $ serverless ???? $ serverless create !--template aws-go !!==> serverless.yml
SAM(-CLI) https://github.com/lambci/docker-lambda Docker images and test runners that replicate the
live AWS Lambda environment
TEMPLATES https://github.com/serverless/examples
serverless.yml functions: post: !<<: *default handler: bin/post events: - http:
path: blog/{slug} method: get
post.go package main import ( "github.com/aws/aws-lambda-go/events" "github.com/aws/aws-lambda-go/lambda" ) func Handler(ctx
context.Context, request Request) (events.APIGatewayProxyResponse, error) { !// put code here } func main() { lambda.Start(Handler) }
post.go func Handler(_, request Request) (_, _) { var buf
bytes.Buffer var slug = request.PathParameters["slug"] … }
post.go func Handler(_, request Request) (_, _) { var buf
bytes.Buffer var slug = request.PathParameters["slug"] post !:= lib.NewPost(slug, lambdaF.Config) if err !:= post.Render(&buf, lambdaF.Renderer); err !!= nil { return events.APIGatewayProxyResponse{StatusCode: 404}, err } … }
post.go func Handler(_, request Request) (_, _) { var buf
bytes.Buffer var slug = request.PathParameters["slug"] post !:= lib.NewPost(slug, lambdaF.Config) if err !:= post.Render(&buf, lambdaF.Renderer); err !!= nil { return events.APIGatewayProxyResponse{StatusCode: 404}, err } return lambdaF.Render(&buf) }
REDUCED x250 0.02$ vs 5$
+VAT = 6$!!! REDUCED x300
SERVERLESS UNDER THE MICROSCOPE
LOG $ serverless invoke -f index -l
LOG $ serverless invoke -f index -l START RequestId: <UUID>
Version: $LATEST END RequestId: <UUID> REPORT RequestId: <UUID> Duration: 1.48 ms Billed Duration: 100 ms Memory Size: 1024 MB Max Memory Used: 31 MB
LOG $ serverless invoke -f index -l START RequestId: <UUID>
Version: $LATEST END RequestId: <UUID> REPORT RequestId: <UUID> Duration: 1.48 ms Billed Duration: 100 ms Memory Size: 1024 MB Max Memory Used: 31 MB
LOG $ serverless invoke -f index -l START RequestId: <UUID>
Version: $LATEST END RequestId: <UUID> REPORT RequestId: <UUID> Duration: 1.48 ms Billed Duration: 100 ms Memory Size: 1024 MB Max Memory Used: 31 MB
WHAT ABOUT CPU?
AWS Lambda allocates CPU power proportional to the memory… (с)
Amazon Docs
AWS Lambda allocates CPU power proportional to the memory… For
example, if you allocate 256 MB memory, your Lambda function will receive twice the CPU share than if you allocated only 128 MB. (с) Amazon Docs
https://aws.amazon.com/lambda/pricing/
6$ / PRICE(100ms) = 28.8M!
MEMORY / CST http://bit.ly/2UZ7xDL Completely Fair Scheduler (CFS)
FREE!
CST https://amzn.to/2Ix2t2O
COST !!<=> LANGUAGE
$ wrk -t10 -c400 -d20s -R40000 http://endpoint.url http://github.com/ssnickolay/dump2019
$ wrk -t10 -c400 -d20s -R40000 $ curl http://endpoint.url/[ruby/go] !=>
{“success”: true} http://github.com/ssnickolay/dump2019
0 250 500 750 1000 0 1 2 4 5
MAX AVG AVG ~80ms BATTLE (128 MB)
0 250 500 750 1000 0 1 2 4 5
MAX AVG AVG ~80ms 31712 requests in 20.01s Requests/sec: 1584.83 BATTLE (128 MB)
0 250 500 750 1000 0 1 2 4 5
0 250 500 750 1000 0 1 2 4 5 MAX AVG AVG ~1.8ms AVG ~80ms 31712 requests in 20.01s Requests/sec: 1584.83 50347 requests in 20.00s Requests/sec: 2517.30 BATTLE (128 MB)
0 250 500 750 1000 0 1 2 4 5
0 250 500 750 1000 0 1 2 4 5 MAX AVG AVG ~1.8ms AVG ~80ms 31712 requests in 20.01s Requests/sec: 1584.83 50347 requests in 20.00s Requests/sec: 2517.30 W IN BATTLE (128 MB)
0 250 500 750 1000 0 1 2 4 5
0 250 500 750 1000 0 1 2 4 5 MAX AVG AVG ~1.8ms AVG ~80ms 31712 requests in 20.01s Requests/sec: 1584.83 50347 requests in 20.00s Requests/sec: 2517.30 W IN + BATTLE (128 MB)
0 250 500 750 1000 0 1 2 4 5
0 250 500 750 1000 0 1 2 4 5 MAX AVG AVG ~1.8ms AVG ~80ms 31712 requests in 20.01s Requests/sec: 1584.83 50347 requests in 20.00s Requests/sec: 2517.30 BATTLE (128 MB)
BATTLE (128 MB) AVG ~1.8ms AVG ~80ms “Billed Duration: 100
ms” (c) Lambda Log
COST !!<=> PACKAGE SIZE
5.4MB 140KB COST !!<=> PACKAGE SIZE MIN SIZE
Go + WRK2 Milliseconds (AVG) 1,7 1,775 1,85 1,925 2
Megabytes 6 25 150 220
5.4MB 140KB COST !!<=> PACKAGE SIZE
http://rubyonjets.com/
COST !!<=> DEPENDENCIES
MINIMIZE THE COMPLEXITY OF YOUR DEPENDENCIES
MINIMIZE THE COMPLEXITY OF YOUR DEPENDENCIES
MINIMIZE THE COMPLEXITY OF YOUR DEPENDENCIES
WAIT A MINUTE
https://stackoverflow.com/q/54110697
When you write your Lambda function code, do not assume
that AWS Lambda automatically reuses the execution context for subsequent function invocations…. (c) AWS Docs AWS LAMBDA EXECUTION CONTEXT
func Fibonacci(n uint) uint { if n !<= 1 {
return n } return Fibonacci(n-1) + Fibonacci(n-2) } OPTION 1
func Fibonacci(n uint) uint { if n !<= 1 {
return n } return Fibonacci(n-1) + Fibonacci(n-2) } func main() { lambda.Start(Handler) } OPTION 1
func Fibonacci(n uint) uint { if n !<= 1 {
return n } return Fibonacci(n-1) + Fibonacci(n-2) } !// Handler is our lambda handler invoked by !// the `lambda.Start` function call func Handler(ctx context.Context) (Response, error) { Fib !:= Fibonacci(33) !// noise here resp !:= Response{ Body: Fib } … } func main() { lambda.Start(Handler) } OPTION 1
Milliseconds (AVG) 0 250 500 750 1000 1 2 3
4 Golang (128MB + Fib) ~ 400ms Option1
var Fib uint func Fibonacci(n uint) uint {…} func Handler(ctx
context.Context) (Response, error) { … } func init() { Fib = Fibonacci(33) !// noise here } func main() {…} OPTION 2
Milliseconds (AVG) 0 250 500 750 1000 1 2 3
4 Golang (128MB + Fib) ~ 400ms Option1 Option2
COST SPEED ⏫
0 15000 30000 45000 60000 0 1 2 4 5
50347 requests in 20.00s Requests/sec: 2517.30 LOAD !!<=> $$$
0 15000 30000 45000 60000 0 1 2 4 5
50347 requests in 20.00s Requests/sec: 2517.30 PROTECTION 1: X-API-KEY
0 15000 30000 45000 60000 0 1 2 4 5
PROTECTION 2: AWS SHIELD 50347 requests in 20.00s Requests/sec: 2517.30
0 15000 30000 45000 60000 0 1 2 4 5
50347 requests in 20.00s Requests/sec: 2517.30 62808 requests in 20.01s Non-2xx or 3xx responses: 62127 1 1000 CONCURRENCY PROTECTION 3: THROTTLING
PREDICTABLY
PREDICTABLY
PREDICTABLY
RECIPE
1. TAKE SERVERLESS
1. TAKE SERVERLESS 2. FOLLOW RULES
1. TAKE SERVERLESS 2. FOLLOW RULES 3. ????
1. TAKE SERVERLESS 2. FOLLOW RULES 3. ???? 4. PROFIT
WHAT ABOUT PRODUCTION?
WHAT ABOUT PRODUCTION? Tung Nguyen
@ssnickolay Nikolay Sverchkov @ssnickolay THANK YOU! @evilmartians http://evl.ms/blog http://evl.ms/telegram
More Decks by Nikolay Sverchkov
See All by Nikolay Sverchkov
Authorization in the GraphQL era
ssnickolay
0
1.2k
Rails 6: ActiveRecord
ssnickolay
0
58
Serverless Is Ruby Future
ssnickolay
0
980
Anything new Rails 6?
ssnickolay
1
1.6k
Evolution of Rails application architecture: 14 years in production
ssnickolay
3
1.6k
A Healthy Monolith
ssnickolay
0
1.7k
Elixir. There and Back Again
ssnickolay
0
940
Other Decks in Programming
See All in Programming
どうしても Eloquent でリポジトリパターンをやりたいんだ #phpcon_niigata / phpcon-niigata-2026
shogogg
0
620
Vapor Mode でアクセシビリティは 壊れないか検証した話
fuqda
1
350
Snowflake Openflow Connector利用の勘所 / 20261024developersio2026tokyo-snowflake-openflow
kasacchiful
PRO
0
360
Growing PicoRuby's Ecosystem on ESP32
yuuu
2
150
선언형 햅틱 라이브러리 진동을 소개합니다
l2hyunwoo
1
320
クラウドゲーミングのパフォーマンス最適化に挑む / Cloud Gaming Performance Tuning
vivion
0
360
将来の標準フォーマッター?ktfmtを深堀る
t45k
0
220
Materialが消えた日 〜真っ白なFlutterキャンバスにRenderObjectで絵を描く〜
umigishiaoi
0
170
動画プレイヤーはなぜカクつくのか 〜Flutterのレンダリングとネイティブの境界を解剖する〜
naruogram
0
340
社内AIエージェント「Askワンバン」における コストチューニングの取り組み
t0yohei
0
170
Engineering Biometric UX Across Android Versions | DroidKnights 2026
ryunen344
1
210
AIが生成したコードは読むべきか?抽象化設計とコードの品質保証 / abstraction-design-and-quality
minodriven
16
9.4k
Featured
See All Featured
What does AI have to do with Human Rights?
axbom
PRO
1
2.5k
SEO for Brand Visibility & Recognition
aleyda
0
4.9k
A brief & incomplete history of UX Design for the World Wide Web: 1989–2019
jct
2
570
sira's awesome portfolio website redesign presentation
elsirapls
0
510
Tell your own story through comics
letsgokoyo
1
1.1k
Chasing Engaging Ingredients in Design
codingconduct
0
400
Principles of Awesome APIs and How to Build Them.
keavy
128
18k
Navigating the moral maze — ethical principles for Al-driven product design
skipperchong
2
690
Why Mistakes Are the Best Teachers: Turning Failure into a Pathway for Growth
auna
0
360
The Cost Of JavaScript in 2023
addyosmani
56
10k
How to Build an AI Search Optimization Roadmap - Criteria and Steps to Take #SEOIRL
aleyda
1
2.3k
Design of three-dimensional binary manipulators for pick-and-place task avoiding obstacles (IECON2024)
konakalab
0
670