https://github.com/mcollina/the-cost-of-logging

My talk "The Cost of Logging" about our uber-fast Pino logger
https://github.com/mcollina/the-cost-of-logging

Last synced: 12 days ago
JSON representation

My talk "The Cost of Logging" about our uber-fast Pino logger

• Host: GitHub
• URL: https://github.com/mcollina/the-cost-of-logging
• Owner: mcollina
• License: mit
• Created: 2016-07-14T13:53:13.000Z (over 8 years ago)
• Default Branch: master
• Last Pushed: 2017-12-12T19:40:44.000Z (about 7 years ago)
• Last Synced: 2024-04-14T16:14:05.078Z (9 months ago)
• Language: HTML
• Homepage: https://mcollina.github.io/the-cost-of-logging/
• Size: 94.1 MB
• Stars: 20
• Watchers: 6
• Forks: 1
• Open Issues: 1
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# The Cost of Logging

## How we realised 

## Why we did it

* talk about NAP 

* Prior to pino, options weren't great

## Logging is hard

* Don't roll your own logger!

* Performance wise

* Architecturally

* Noise filtering

## Adapters of adapters of adapters

* Too many layers

* Too much complexity

* To high a cost of performance

## Newline Delimited JSON

* Everyone knows how to work with JSON

* It has become the logging standard

## Introducing Pino

* High speed logging - 10000 logs in 250ms

* Bunyan compatible interface

* Faster than other JSON loggers (winston, bunyan)

* Faster even, then non-JSON low-feature loggers (mithril, debug)

## Why Pino is fast

* not calling JSON.stringify

* flatstr

* fast-safe-stringify

* quick-format

## Making Pino even faster

* Extreme mode

* 10000 logs in 100ms

* 4k batching

## How we did it

* 0x

* tracing

* experience

* use common modules, then rewrite when they become bottleneck  

## Pino Philosophy

* Pino provides both high speed logging *and* recommends an approach to logging related actions

* Log rotation

* Transports

* Post-formatting (date, spacing, etc.)

## Server integration

* hapi-pino (code example)

* express-pino-logger (code example)

* restify-pino-logger (code example)

* koa-pino-logger (code example)

* pino-http (code example)

## ELK / Cloudwatch integration

* todo

## Ecosystem impact

* usage is growing

* apps/processes are speeding up essentially for free

* inspiring other projects (rvagg and bole)

* knowledge osmosis - approaches in pino and tools used for pino are being used in other projects (fast-safe-stringify hapi etc.)