Skip to content

Instantly share code, notes, and snippets.

@tuliren

tuliren/tammy-butow-scaling-to-millions-of-users.md

Last active November 16, 2018 19:05
Show Gist options
  • Save tuliren/faf0e125e84c63a8467d189750395a1e to your computer and use it in GitHub Desktop.
Save tuliren/faf0e125e84c63a8467d189750395a1e to your computer and use it in GitHub Desktop.
Download ZIP
Raw
tammy-butow-scaling-to-millions-of-users.md

Lessons from Scaling to Hundreds of Millions of Users

Tammy Butow presentation on DigitalOcean Tide SF

Before launch

  • Kubernetes
    • 3 primaries and 3 nodes
    • Docker containers
    • Load balancer
    • Private network
  • Sharded MySQL
    • Percona community
    • Semi-sync replication
  • Engineering tools
    • Git
    • CI (Circle CI)
    • Infra automation (Terraform)
  • Backups
    • Percona backups, short and long

With 5 enterprise customers

  • Small data
  • Specific infra

With 3 million users

  • Caching
    • Memcache
  • Big data
    • Hadoop, Spark, Pig
  • Moar specific infra

With 400 million users

  • Distibuted datastore
    • Built in-house
  • Moar specific infra

Summary

  • Start from simple components
  • Build new layers on top on existing layers
    • Because migration can be too costly and not always possible
  • Prioritise reliability, durability, and performance
  • Focus on making sure "it just works"
  • Infra and product engineering work together
  • Sweat the details and aim higher each day
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment