Deploy Skill

Guide for deploying the Site Scanning Engine to Cloud.gov (Cloud Foundry).

Important Notes

• Deployments are handled by GitHub Actions workflows, not a local script or manual cf push
• Use cf CLI for monitoring and troubleshooting a deployed environment. Information on the Cloudfoundry CLI can be found at https://docs.cloudfoundry.org/cf-cli/. General advice on deployment with Cloudfoundry is at https://docs.cloudfoundry.org/devguide/deploy-apps/deploy-app.html
• Web dashboard available for monitoring logs, but CLI is preferred for actual work

Prerequisites

• Cloud Foundry CLI (cf) installed (for monitoring/troubleshooting)
• Cloud.gov account with appropriate permissions
• Access to the correct Cloud.gov space

Deployment Process

Deployments are triggered via GitHub Actions:

• deploy.yml deploys to the prod space using manifest.yml + vars-prod.yml. Triggered manually only (workflow_dispatch).
• dev.yml deploys to the dev space using manifest.yml + vars-dev.yml. Runs automatically on pull requests to main, and can also be triggered manually.

To trigger a deployment, run the workflow from the GitHub Actions tab (or gh workflow run deploy.yml / gh workflow run dev.yml). Each workflow installs dependencies, builds all apps, then runs cf push -f manifest.yml --vars-file <vars file> using the CF_USERNAME / CF_PASSWORD GitHub secrets.

1. Login to Cloud.gov (for monitoring/troubleshooting)

cf login -a api.fr.cloud.gov --sso

This will:

1. Open browser for SSO authentication
2. Provide a temporary code
3. Enter code in terminal to complete login

2. Verify Target Space

cf target

Ensure you're targeting the correct org and space.

Configuration Files

manifest.yml

Main Cloud Foundry manifest - defines:

• Application name
• Memory allocation
• Worker counts (7 scan-engine workers in production)
• Buildpacks
• Services to bind

vars-prod.yml / vars-dev.yml

Environment-specific variables, passed to cf push via --vars-file:

• Instance counts, memory, disk
• Scan/snapshot schedules

vars.yml (symlink)

Points to environment-specific variables:

• vars.yml → vars-prod.yml (production)

To change a resource setting (e.g., worker count or memory), edit the relevant vars-*.yml file and deploy via the workflow above -- manual cf scale changes will be reverted on the next deploy.

vars.yml (symlink)

Points to environment-specific variables:

• vars.yml → vars-prod.yml (production)
• vars-dev.yml (development)

Post-Deployment Verification

Check Application Status

cf apps

Look for (instance counts come from vars-prod.yml / vars-dev.yml):

• site-scanner-api
• site-scanner-consumer - 7 instances in production

View Recent Logs

cf logs <app-name> --recent

Example:

cf logs site-scanner-api --recent
cf logs site-scanner-consumer --recent

Stream Live Logs

cf logs <app-name>

Press Ctrl+C to stop streaming.

Managing Services

List Bound Services

cf services

Shows:

• PostgreSQL database
• Redis queue
• S3 bucket

View Environment Variables

cf env <app-name>

Example:

cf env site-scanner-api

Restart Application

cf restart <app-name>

Restage Application (rebuild)

cf restage <app-name>

Use after:

• Buildpack updates
• Service binding changes

Scaling

Scale Instances

cf scale <app-name> -i <count>

Example:

cf scale site-scanner-consumer -i 7

Scale Memory

cf scale <app-name> -m <memory>

Example:

cf scale site-scanner-api -m 1G

Troubleshooting

Application Won't Start

1. Check recent logs: cf logs <app-name> --recent
2. Verify services are bound: cf services
3. Check environment variables: cf env <app-name>
4. Review manifest.yml for configuration errors

Connection Issues

• Verify PostgreSQL service is running
• Check Redis service status
• Verify the bound services have the expected credentials: cf env <app-name>

Deployment Fails

• Check the failed step in the GitHub Actions run for deploy.yml / dev.yml
• Check Cloud.gov status page
• Verify you have sufficient quota: cf quotas
• Review buildpack compatibility

Important Reminders

• Docker is only for local development - not used in Cloud Foundry
• API must deploy successfully before scan-engine (database migrations)
• Always verify services are bound after deployment
• Keep vars.yml symlink pointing to correct environment file
• Resource changes (memory, instance counts) belong in vars-prod.yml / vars-dev.yml, not ad-hoc cf scale calls, since the next deploy will reset them