What is Terraform used for in Amazon Connect environments?
Infrastructure as Code (IaC) tool that deploys and manages AWS resources through code instead of manual console work. Manages Lambdas | DynamoDB tables | IAM roles | dashboards | and all other infrastructure.
Emphasize that Terraform prevents configuration drift (manual changes that don’t match documentation), enables infrastructure versioning through Git, and makes it possible to spin up identical environments quickly - critical for scaling contact center operations.
How are Lambda changes deployed in enterprise environments?
Through Git repositories and CI/CD pipelines. Create a feature branch | commit changes | run the pipeline which deploys via Terraform. Should NOT make changes directly in the AWS console.
What is AWS CDK?
Cloud Development Kit - an alternative to Terraform for Infrastructure as Code. Know what it is even if you primarily use Terraform so you can reference it in interviews.
How would you describe Terraform experience in an interview?
“I took it upon myself to automate console tasks by working with DevOps to create Terraform configurations for pipeline deployments. I know enough to read Terraform files | understand them | and make small changes when required.”
What does a typical Terraform deployment pipeline look like?
Script runs | generates Terraform configuration | pipeline executes | grabs config and deploys to AWS environment.
Why is Terraform knowledge a big plus for Amazon Connect roles?
You don’t need to be a full Terraform developer but knowing basic definitions and concepts shows you understand enterprise-level deployment practices and automation.
What is a static reference in an Amazon Connect contact flow?
A hardcoded reference to a specific resource like a queue name | Lambda ARN | or prompt that is explicitly set in the flow block configuration.
What is a dynamic reference in an Amazon Connect contact flow?
A reference that pulls its value from an attribute or variable at runtime. Allows the same flow block to point to different resources based on context using syntax like $.Attributes.QueueName instead of hardcoding.
Why would a flow block show No Match in the console?
When using dynamic references the referenced attribute or variable doesn’t resolve to a valid resource. The actual resource is determined at runtime not design time.
What is the advantage of dynamic references over static?
Flexibility and reusability. One contact flow can handle multiple scenarios without duplicating flows. Change behavior by updating attributes rather than modifying the flow itself.
Where do dynamic reference values typically come from?
Contact attributes set earlier in the flow | Lambda function returns | DynamoDB lookups | or external system integrations.
What are the two architecture options when creating a Lambda?
x86_64 and ARM64.
Which Lambda architecture should you typically choose and why?
x86_64 because it has better compatibility with machine learning libraries and dependencies. ARM64 can work but x86 is the safer default.
What is Amazon Connect’s default timeout for Lambda invocations?
8 seconds.
What should you set your Lambda timeout to when used with Amazon Connect?
8 seconds or less to match Amazon Connect’s timeout limit.
What happens if your Lambda exceeds the Amazon Connect timeout?
The contact flow hits the error branch even if the Lambda would eventually complete successfully. The caller experiences a failure.
Why is the 8 second timeout important for Lambda design?
You must design efficient code that completes within this window. If operations take longer you need to optimize or consider alternative architectures like async processing.
Does increasing Lambda memory increase execution speed?
Yes and No. AWS increases CPU power proportionally with memory so more memory equals more speed. But if your Lambda only needs 100MB setting it to 10GB won’t make it faster because it’s not memory bottlenecked.
How do you know if your Lambda needs more memory?
Check CloudWatch logs which show actual memory used per invocation. If usage consistently hits your allocated limit like using 128MB when set to 128MB then increase it.
What is the recommended approach for scaling Lambda memory?
Start low | monitor usage | scale incrementally from 128MB to 256MB to 512MB and so on. Only increase when the Lambda is hitting its memory ceiling.
How does Lambda memory affect billing?
You are charged for ALLOCATED memory not just what is used. Setting 10GB when you only need 100MB wastes money on every invocation.
What is Lambda ephemeral storage?
Temporary disk storage like RAM available during execution. Used for processing files or temporary data that doesn’t need to persist.
What is a VPC?
Virtual Private Cloud. A virtual network in AWS that isolates your resources similar to a local on-premises network but in the cloud.
Should Lambdas be in a VPC in production?
Yes always use VPC in production for an extra layer of security. It’s not hard to set up and adds significant protection.