Problem

DNS management for ai.coder.com and proxy domains currently requires manual Slack requests to #help-me-ops for CloudFlare changes. This creates friction during workshops and incident response, and is a potential single point of failure.

Context

Domains Managed in CloudFlare:

• ai.coder.com +*.ai.coder.com → us-east-2 NLB (Control Plane)
• oregon-proxy.ai.coder.com +*.oregon-proxy.ai.coder.com → us-west-2 NLB
• emea-proxy.ai.coder.com +*.emea-proxy.ai.coder.com → eu-west-2 NLB

Current Process:

1. Request DNS change in #help-me-ops Slack channel
2. Wait for ops team response
3. Manual CloudFlare console changes
4. Verify propagation

Pain Points:

• Manual process doesn't scale during incidents
• No self-service for infrastructure team
• Dependency on ops team availability
• No automated validation of DNS configuration

Requirements

Terraform/IaC Management

• Migrate CloudFlare DNS records to Terraform
• Use CloudFlare Terraform provider
• Store state in shared backend (S3)
• Document DNS change process via Infrastructure as Code
• Enable self-service DNS changes via PR/approval workflow

Automated Validation

• Add DNS validation to pre-workshop checklist
• Implement automated tests:

  # Verify all 6 domains resolve correctlydig ai.coder.comdig oregon-proxy.ai.coder.com  dig emea-proxy.ai.coder.com# Verify wildcard subdomainsdig test.ai.coder.comdig test.oregon-proxy.ai.coder.comdig test.emea-proxy.ai.coder.com

• Run validation as part of CI/CD pipeline
• Alert on DNS misconfiguration

Documentation

• Document CloudFlare API access requirements
• Document Terraform workflow for DNS changes
• Add DNS troubleshooting to incident runbook (Create pre-workshop validation checklist and runbook #4) ✅ (already added)
• Create rollback procedure for DNS changes

Self-Service Workflow

• Grant infrastructure team CloudFlare API access (scoped to ai.coder.com zone)
• Implement PR-based approval workflow:
  1. Create Terraform change in PR
  2. Automated validation/plan
  3. Team review
  4. Apply after approval
• Set up notifications for DNS changes (Slack, email)

Monitoring & Alerting

• Monitor DNS resolution for all 6 domains
• Alert on:
  • DNS resolution failures
  • TTL expiration
  • Certificate expiration (related to DNS)
  • Unexpected DNS changes
• Add to monitoring dashboard (Implement comprehensive resource monitoring and alerting #6)

Success Criteria

• DNS changes can be made via Terraform without manual Slack requests
• Infrastructure team has self-service access to CloudFlare DNS
• DNS configuration validated automatically before and after workshops
• DNS issues detected and alerted before user impact
• Zero workshop delays due to DNS misconfigurations

Implementation Notes

CloudFlare Terraform Example:

resource"cloudflare_record""ai_coder_com" {zone_id=var.cloudflare_zone_idname="ai"value=aws_lb.coder_nlb_us_east_2.dns_nametype="CNAME"ttl=300proxied=false}resource"cloudflare_record""ai_coder_com_wildcard" {zone_id=var.cloudflare_zone_idname="*.ai"value=aws_lb.coder_nlb_us_east_2.dns_nametype="CNAME"ttl=300proxied=false}resource"cloudflare_record""oregon_proxy" {zone_id=var.cloudflare_zone_idname="oregon-proxy.ai"value=aws_lb.coder_nlb_us_west_2.dns_nametype="CNAME"ttl=300proxied=false}# ... additional records for London proxy and wildcards

CloudFlare API Scoping:

• Use API token (not Global API Key)
• Scope toai.coder.com zone only
• Grant DNS edit permissions only
• Rotate token periodically

Security Considerations

• CloudFlare API token stored in secure secret management (AWS Secrets Manager, HashiCorp Vault)
• API token scoped to minimum required permissions
• Audit log for all DNS changes
• Require PR approval for production DNS changes

Future Domains

This infrastructure should support upcoming domains:

• coderdemo.io - SE official demo environment
• devcoder.io - CS / Engineering collaboration environment

Related

Sept 30 Workshop Postmortem
#2 (Image management and subdomain routing)
#4 (Pre-workshop validation checklist)
#6 (Monitoring and alerting)
Incident Runbook - Subdomain Routing Failures
Pre-Workshop Checklist - CloudFlare DNS Verification