Web tools for Age-Period-Cohort Analysis

Age Period Cohort (APC) Web Tool

Age-Period-Cohort analysis identifies patterns in cancer incidence or mortality rates from population-based Count (numerator) and Population (denominator) data. Often the data come from a Cancer Registry (e.g., SEER) in the form of a table showing the numbers of cancer cases or cancer deaths (counts) and corresponding person-years at risk (population) for particular age groups and calendar time periods. This toolset provides a comprehensive solution to age-period-cohort analysis for cancer endpoints in defined populations and time periods.

Comparative Age Period Cohort (CrossTalk) Web Tool

---

CI/CD Process

Overview

This project uses GitHub Actions to automate the deployment of the APC web tools to AWS ECS Fargate. The CI/CD pipeline is manually triggered via workflow dispatch and supports deployments to multiple environments.

Workflow Trigger

The deployment workflow (.github/workflows/apc-fargate-deploy.yml) is triggered manually via the GitHub Actions UI with the following options:

• Tier Selection: Choose the target environment (dev, qa, stage, prod)

Environment Mapping

The CI/CD pipeline supports four deployment tiers with the following configurations:

Tier	Environment	Image Tag	AWS Account	Description
dev	Development	development-backend-*	Configured via secrets	Development environment for testing new features
qa	QA	development-backend-*	Configured via secrets	Quality assurance testing environment
stage	Staging	release-backend-*	Configured via secrets	Pre-production staging environment
prod	Production	release-backend-*	Configured via secrets	Production environment

Image Tier Logic:

• dev and qa tiers use development tagged images
• stage and prod tiers use release tagged images

Deployment Process

The deployment workflow performs the following steps:

1. Authentication

   • Assumes AWS IAM role using OIDC
   • Session name includes tier, app name, and branch reference

2. Build Configuration

   • Sets environment variables including:
     • Branch name and Git tag
     • Timestamp for versioning
     • ECR repository URLs
     • Backend image tags (versioned and latest)

3. AWS Parameter Retrieval

   • Fetches deployment configuration from AWS Systems Manager Parameter Store:
     • ECS cluster name
     • Task definition name
     • ECS service name
     • Task execution and task role ARN

4. Docker Build

   • Uses Docker Buildx for multi-platform builds
   • Builds backend image from ./apc/docker/backend.dockerfile
   • Tags images with both versioned tag and latest tag
   • Pushes to Amazon ECR with build cache optimization

5. Task Definition

   • Renders ECS task definition from template (.github/aws/web.yml)
   • Substitutes environment-specific variables using envsubst
   • Registers new task definition revision with AWS ECS

6. Service Deployment

   • Updates ECS service with new task definition
   • Forces new deployment to ensure latest image is used
   • Sets desired count to 1
   • Propagates tags from task definition

Infrastructure Components

ECS Fargate Configuration:

• CPU Units: 2048 (2 vCPU)
• Memory: 4096 MB (4 GB)
• Network Mode: awsvpc
• Backend Container Port: 80

Containers:

1. Backend Container

   • Image: Latest versioned backend image from ECR
   • Environment: Production Flask environment
   • Logging: AWS FireLens with Datadog integration

2. Logs Container

   • Image: AWS public fluent-bit image
   • Purpose: Log routing via FireLens

Required AWS Resources

Each tier requires the following AWS resources to be pre-configured:

AWS Systems Manager Parameters:

• ECS cluster configuration
• Task definition settings
• Service configuration
• IAM role ARNs

Logging Integration:

• CloudWatch log groups per tier
• Datadog configuration (endpoint and API key stored as SecureString parameters)

IAM Roles:

• GitHub Actions CICD role with OIDC trust relationship
• ECS task execution role with ECR and CloudWatch permissions
• ECS task role with application-specific permissions

Amazon ECR:

• Container image repository in us-east-1 region

Deployment Commands

To deploy to a specific environment:

1. Navigate to the Actions tab in GitHub
2. Select the APC Fargate Deploy workflow
3. Click Run workflow
4. Select the desired tier (dev/qa/stage/prod)
5. Click Run workflow button

Monitoring

• CloudWatch Logs: /analysistools/{tier}/apc/web
• Datadog: Logs tagTier-specific log groups
• Datadog: Logs tagged with project and tier informationloyments