Exprimir cada centavo - DevOpsConf 2015

Transcript

1. Exprimir cada centavo
2. @sebamontini Operations Manager @ Nubity Amateur golfer, whisky enthusiast and cloud evangelist
3. Current Generation T2 - General Purpose M4/M3 - General Purpose (Haswell+AVX2/Ivy Bridge) C4/C3 - Compute Optimized (Haswell+AVX2/Ivy Bridge) High performance Applications R3 - Memory Optimized (Ivy Bridge) Distributed Database & Cache, Hungry-hungry Java G2 - GPU Optimized (Sandy Bridge + CUDA) 3D video streaming, machine learning, video encoding, little coins I2 - High I/O (Ivy Bridge) Cassandra, MongoDB, Hadoop D2 - Dense-Storage (Haswell) Map Reduce, MPP, Warehouse Instance Types
4. Instance Sizes Small Instances No EBS optimization Low Network Performance Small local instance storage Medium Instances EBS optimization (varies) Moderate Network Performance Moderate local instance storage Large Instances EBS optimization (varies) 10gb Network Connection SSD local instance storage
5. T2 Instances Family The T2 family of instance types could be a good option for development and test workloads, especially continuous integration (CI) and automated testing. T2 instance types are also a fit for single-server workloads where the server sits idle much of the time. Instance type Initial CPU credit* CPU credits earned per hour Base performance (CPU utilization) Maximum earned CPU credit balance Cores RAM (GB) Cost/Hour Cost/Month t2.micro 30 6 10% 144 1 1 0.013 9.36 t2.small 30 12 20% 288 1 2 0.026 18.72 t2.medium 60 24 40%** 576 2 4 0.052 37.44 t2.large 60 36 60%** 864 2 8 0.104 74.88
6. Reserved Instances
7. Spot Instances Instance Type On Demand Spot m3.medium $0.067 $0.0081 12.09% m3.large $0.133 $0.0201 15.11% m3.xlarge $0.266 $0.0387 14.55% m3.2xlarge $0.532 $0.0708 13.31% m4.large $0.126 $0.0137 10.87% m4.xlarge $0.252 $0.0267 10.60% m4.2xlarge $0.504 $0.0521 10.34% m4.4xlarge $1.008 $0.1043 10.35% m4.10xlarge $2.520 $0.3285 13.04%
8. Spot Bid Advisor
9. Spot Instances New Resource-Oriented Bidding EC2 will use this information to build the fleet using the most economical combination of available instance types, looking for instances that have the lowest Spot price per unit. 2 x r3.8xlarge 4 x r3.4xlarge 8 x r3.2xlarge 16 x r3.xlarge 32 x r3.large Or something more complex and heterogeneous, such as: 1 x r3.8xlarge and 2 x r3.4xlarge 2 x r3.4xlarge and 8 x r3.xlarge 8 x r3.xlarge and 16 x r3.large
10. S3 & Glacier Glacier Pricing Model: • 1Tb = $ 7.17 • 1Tb retrival in 1 month = $ 9.73 • 1Tb retrival in 1 day = $ 306.99 • 1Tb retrival in 4hs = $ 1840
11. Pitfalls Undersizing your servers - Build T1 servers that will spawn alarms constantly - Jenkins build time up to the roof - Rebuilding servers Oversizing your servers - Wasted CPU/Memory processing & storage laying around - RDS storage that will eat you alive - Rebuilding servers API hits - R53 TTL & Health check nightmares - Extreme monitoring gone wrong - Read. The. Fine. Print.
12. Pitfalls Know your infrastructure “Hey! That team migration was amazing! Can’t we just do the same and move our entire storage to AWS?” Understand your applications “So, I’m thinking about moving Project X to AWS. It should be super simple!! It’s just a couple of APIs!” Make your team know & understand that most migrations are an architectural challenge “Can it be done by next sprint?”
13. Forecasting http://calculator.s3.amazonaws.com/index.html
14. is it worth it?
15. Preguntas?
16. Gracias! Sebastian Montini smontini@nubity.com @sebamontini