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
Monitoring on a budget
Search
C J Silverio
June 03, 2016
Programming
2
280
Monitoring on a budget
How npm does monitoring & metrics on a budget.
C J Silverio
June 03, 2016
Tweet
Share
More Decks by C J Silverio
See All by C J Silverio
The economics of package management
ceejbot
4
1.5k
The future of (javascript) modules (in node)
ceejbot
1
270
Keeping JavaScript safe
ceejbot
3
400
ceej's how to solve it
ceejbot
6
760
work-life balance at npm
ceejbot
5
780
hash functions and you!
ceejbot
2
340
The accidental noder
ceejbot
2
140
Design Patterns & Modularity in the npm Registry
ceejbot
3
180
Cheating Gall's Law: MediterraneaJS edition
ceejbot
4
320
Other Decks in Programming
See All in Programming
Universal Linksの実装方法と陥りがちな罠
kaitokudou
1
220
Tuning GraphQL on Rails
pyama86
2
1k
シールドクラスをはじめよう / Getting Started with Sealed Classes
mackey0225
3
400
アジャイルを支えるテストアーキテクチャ設計/Test Architecting for Agile
goyoki
7
2.8k
約9000個の自動テストの 時間を50分->10分に短縮 Flakyテストを1%以下に抑えた話
hatsu38
23
11k
ピラミッド、アイスクリームコーン、SMURF: 自動テストの最適バランスを求めて / Pyramid Ice-Cream-Cone and SMURF
twada
PRO
9
990
とにかくAWS GameDay!AWSは世界の共通言語! / Anyway, AWS GameDay! AWS is the world's lingua franca!
seike460
PRO
1
540
CSC509 Lecture 08
javiergs
PRO
0
110
開発効率向上のためのリファクタリングの一歩目の選択肢 ~コード分割~ / JJUG CCC 2024 Fall
ryounasso
0
360
ECS Service Connectのこれまでのアップデートと今後のRoadmapを見てみる
tkikuc
2
210
Vue3の一歩踏み込んだパフォーマンスチューニング2024
hal_spidernight
3
3.1k
Go言語でターミナルフレンドリーなAIコマンド、afaを作った/fukuokago20_afa
monochromegane
2
140
Featured
See All Featured
CSS Pre-Processors: Stylus, Less & Sass
bermonpainter
355
29k
YesSQL, Process and Tooling at Scale
rocio
167
14k
Unsuck your backbone
ammeep
668
57k
Build your cross-platform service in a week with App Engine
jlugia
229
18k
Teambox: Starting and Learning
jrom
132
8.7k
Refactoring Trust on Your Teams (GOTO; Chicago 2020)
rmw
31
2.7k
Designing for Performance
lara
604
68k
What's in a price? How to price your products and services
michaelherold
243
12k
How to Think Like a Performance Engineer
csswizardry
19
1.1k
Helping Users Find Their Own Way: Creating Modern Search Experiences
danielanewman
29
2.2k
The Psychology of Web Performance [Beyond Tellerrand 2023]
tammyeverts
41
2.1k
4 Signs Your Business is Dying
shpigford
180
21k
Transcript
monitoring on a budget
a few animated gifs with the Twelfth Doctor (0 cats)
C J Silverio vp of engineering, @ceejbot
let's talk npm by the numbers
205 million packages Tuesday 10K requests/sec
npm is 25 people 4 of us run the registry
when the company was formed 5 people total
you outsource many services when you're tiny
you pull them back in-house when you succeed
success is sometimes a catastrophe
npm's scale: runaway success npm's staff: wouldn't this be neat
mission: know this on a budget
None
2 questions: is the registry up? how well is it
performing?
is the registry up? monitoring
how well is it performing? metrics
monitoring
monitoring == pull ask questions that you know the right
answers for
Is this host up? Is this cert about to expire?
Is the DB replication keeping up?
if you get the wrong answer somebody gets paged
nagios state of the art in free
None
It's okay. We never look at it. It just triggers
Pager Duty.
nagios’s virtues: reliability & custom checks
goal: never page anybody
self-healing checks automate the fix if you can!
monitoring == unit tests a ratchet for continuous improvement
external monitoring ping services
you must monitor but that's just the start
monitoring tells you what it doesn't tell you why
metrics
Q: What's a metric? A: A name + a value
+ a time.
counter: it happened N times gauge: it's Y-sized right now
rate: it's happening N times/second timing: it took X milliseconds
metrics == push the app gives you numbers
emit from a service store in timeseries db query &
graph
the usual stack statsd ➜ graphite ➜ grafana
None
None
statsd uses UDP
Q: Why not send metrics over UDP? A: You care
about receiving them.
just try to install graphite
for-pay/SAAS services exist but I can't afford them
monitoring 400 processes right now 12+ GB of log data
a day
interlude: when should you pay?
convert the £$€ cost into engineer hours/month
pay when it's cheaper than investing an engineer (be honest
about the cost)
numbat was born “How hard can it be?” I said.
None
https://github.com/ numbat-metrics numbat - powered metrics
None
npm’s stack numbat ➜ influxdb ➜ grafana
var Emitter = require('numbat-emitter'); var emitter = new Emitter({ uri:
'tcp://localhost:3333', app: 'www', }); process.emit('metric', { name: 'request.latency', value: 30 }); process.emit('metric', { name: 'disk.used.percent', value: 36 }); process.emit('metric', { name: 'login' });
so easy to emit a metric that we just do
it any time something interesting happens
4000 metrics/sec from the registry
None
None
None
None
metrics ➜ alerts
Server handling expected traffic? Latency higher than normal? Error rate
higher than usual?
metrics comprise a data stream send the stream to more
than one place!
anomaly detection
recap time!
your web apps are backed by something
what's it up to? how do you know?
get data on what your services are up to
what: monitoring yes/no questions
why: metrics data changing over time
next: anomaly detection predictions & trends
automate don't require humans
npm install -g npm@latest @ceejbot on all the things npm
loves you