Pro Yearly is on sale from $80 to $50! »

Scaling Down (Running Large Sites Locally)

Scaling Down (Running Large Sites Locally)

A talk I gave at PyCon AU 2017.

077e9a0cb34fa3eba2699240c9509717?s=128

Andrew Godwin

August 05, 2017
Tweet

Transcript

  1. Andrew Godwin

  2. Hi, I’m Andrew Godwin

  3. 2013

  4. None
  5. None
  6. 2017

  7. None
  8. ~150 Docker containers

  9. ~24GB of RAM

  10. None
  11. docker-compose

  12. docker-compose

  13. bay

  14. None
  15. None
  16. None
  17. Containers & Links

  18. Django Core MySQL Payments Audit MySQL Redis

  19. Django Core MySQL Payments Audit MySQL Redis

  20. Django Core MySQL Payments Audit MySQL Redis

  21. Django Core MySQL Redis “tiny”

  22. Did it boot yet?

  23. Docker containers run only one thing

  24. None
  25. None
  26. None
  27. What if you don’t have a custom init?

  28. None
  29. Rebuilding is Bad

  30. Rebuilding core is about 10 mins

  31. Mounts (“devmodes”) Core mount ➜ Library mount ➜

  32. None
  33. Our dependencies change every few hours

  34. Docker images are portable

  35. Pull rather than build

  36. None
  37. What’s Next?

  38. Automated image building

  39. Running images remotely

  40. Running whole stacks remotely

  41. Trying not to reinvent the wheel

  42. github.com/eventbrite/bay

  43. Thanks. github.com/eventbrite/bay