13. Requirements
Scaling Spreaker
Fast Automatized
New instances up and No human intervention
running in less then 5 mins
14. Auto-Scaling
Scaling Spreaker
TRACK UP / DOWN
ANALYZE
METRICS SCALE
15. Auto-Scaling
Scaling Spreaker
TRACK UP / DOWN
ANALYZE
METRICS SCALE
Zenoss + Logstash Skynet Spreaker + EC2 API
16. Auto-Scaling
Scaling Spreaker
TRACK UP / DOWN
ANALYZE
METRICS SCALE
Zenoss + Logstash Skynet Spreaker + EC2 API
{
You can also use
CloudWatch
17. Auto-Scaling
Scaling Spreaker
1. Split the infrastructure into roles
web
streaming
api
18. Auto-Scaling
Scaling Spreaker
1. Split the infrastructure into roles
db-master
web
gearman
streaming
db-slave
redis
api
recording search
memcache
encoder
mongo
web worker
19. Auto-Scaling
Scaling Spreaker
2. Assign roles to each instance (using EC2 tags)
Instance 1 Instance 2 Instance 3
web web streaming
api api
20. Auto-Scaling
Scaling Spreaker
3. Start roles at instance boot
a script fetch roles,
install code (latest build),
configure and start services
21. Auto-Scaling
Scaling Spreaker
4. Ready to auto-scale
To add another web server,
turn on a new instance with web role