Intro

• Function as a Service (FaaS)
• Serverless
• Auto-scaling
• Pay per request (number of invocations) and compute time
• Free tier: 1,000,000 requests and 400,000 GBs of compute time
• Supported runtime

Performance & Limits

• RAM: 128MB (default) - 10 GB (1 MB increments)
• Disk capacity (/tmp): 512 MB (free) - 10 GB
• Cannot configure vCPU count directly, increase RAM to increase vCPU count
  • RAM = 1,792 MB ⇒ vCPU = 1
  • For vCPU > 1, need to use multi-threading to take advantage of multiple cores
• Timeout: default 3s, max 15 mins (900s)
• Environment variables: max 4 KB
• Deployment
  • Compressed: max 50 MB
  • Uncompressed: max 250 MB (for more use /tmp)

Cold Start

• To minimize partial cold start time, minimize the time taken to execute the code outside the lambda handler by:
  • Reducing the deployment package size
  • Increasing computing power (by increasing the allocated memory)

Access Control