Operations Guide
Last Updated: September 12, 2026
This section provides a production operations guide for EKS Auto Mode environments. It covers infrastructure provisioning with Terraform, CI/CD pipelines, GitOps-based deployment, scaling, observability, resource optimization, and upgrades.
Target Audience
- Platform Engineers building production environments with EKS Auto Mode
- Infrastructure Engineers operating Terraform/Terragrunt-based IaC
- DevOps Engineers building CI/CD pipelines with GitLab CI and ArgoCD
- SREs operating Prometheus, Grafana, and Loki observability stacks
Prerequisites
Before starting this operations guide, ensure familiarity with:
- Getting Started with EKS Auto Mode
- Terraform basics (resources, modules, state management)
- Kubernetes Core Concepts
- kubectl and Helm CLI experience
Table of Contents
| # | Document | Key Topics |
|---|---|---|
| 01 | Terraform 3-Layer Infrastructure Setup | VPC, EKS Auto Mode, Pod Identity with 3-Layer Terraform |
| 02 | NLB Weighted Routing and Blue/Green | Dual cluster architecture, NLB weights, DNS routing |
| 03 | CI Pipelines | ECR, GitLab Runner, GitHub ARC, multi-platform builds |
| 04 | ArgoCD Multi-Cluster | Hub-spoke, ApplicationSet, IAM Identity Center SSO |
| 05 | GitOps Automation | Atlantis, FluxCD, Terraform Cloud, AIOps |
| 06 | Scaling Strategies | HPA custom metrics, KEDA, VPA, Spot utilization |
| 07 | Operational Alert Configuration | Network/CPU/Disk/Auto Mode node termination alerts |
| 08 | Observability Analysis | Logs/Metrics/Traces correlation, PromQL, LogQL, TraceQL |
| 09 | Observability Stack Operations | Loki, Tempo, Prometheus/AMP installation and operations |
| 10 | Resource Optimization | Requests/Limits, JVM tuning, framework-specific guide |
| 11 | EKS Upgrades | Staged Auto Mode upgrade, blue/green strategy |
| 12 | Event Capacity Planning | Traffic events, capacity and recovery planning |
| 13 | FinOps Cost Management | Cost visibility, allocation, optimization |
| 14 | Tekton Pipelines | Kubernetes-native CI pipelines |
| 15 | Zonal Cluster Operations | LB weight shifting and TargetGroupBinding, native rollback, Kafka/Redis/Aurora AZ-affine reads |
| 16 | Troubleshooting Playbook | Symptom → diagnosis → cause → fix: Pending/ImagePull/CrashLoop/NotReady/PVC, IRSA, VPC CNI, Karpenter, kubectl cheat sheet |
| 17 | EKS Spot Production Experiments | Interruption, concurrent reclamation, fallback, result records, SLO/cost assessment, rollback |
Learning Path
Recommended Order
- Infrastructure (01-02): Provision VPC/EKS with Terraform
- CI/CD (03-05): Build pipelines and GitOps deployment
- Scaling (06): Establish scaling strategies for workloads
- Observability (07-09): Build monitoring, alerting, and analysis systems
- Optimization (10): Resource efficiency and cost optimization
- Upgrades (11): Validate availability and recovery procedures
- Specialized operations (12–16): Capacity events, FinOps, Tekton, zonal operations, and troubleshooting
By Role
| Role | Priority Documents |
|---|---|
| Platform Engineer | 01, 02, 04, 11 |
| DevOps Engineer | 03, 04, 05 |
| SRE | 06, 07, 08, 09, 10 |
| Infrastructure Engineer | 01, 02, 11 |
Relationship with Existing Documentation
This operations guide complements existing concept documentation with practical, code-focused guides:
Concept Documentation
- EKS Auto Mode - Architecture and concepts
- ArgoCD - GitOps fundamentals
- KEDA - Event-driven autoscaling concepts
- Karpenter - Node provisioning concepts
This Operations Guide
- Terraform HCL code for production infrastructure
- YAML manifests for Kubernetes resources
- PromQL/LogQL queries for observability
- Bash scripts for operational automation
- Step-by-step procedures with validation
Quick Reference
Common Operations
| Task | Document | Section |
|---|---|---|
| Create new EKS cluster | 01-infrastructure-setup | 3-Layer Terraform |
| Add new application to ArgoCD | 04-gitops-multi-cluster | ApplicationSet |
| Configure HPA with custom metrics | 06-scaling-strategies | Custom Metrics |
| Set up alerting for new service | 07-observability-alerts | Alert Rules |
| Upgrade EKS version | 11-upgrade-operations | Auto Mode Upgrade |
Emergency Procedures
| Scenario | Document | Section |
|---|---|---|
| Rollback deployment | 04-gitops-multi-cluster | Rollback |
| Rollback EKS upgrade | 11-upgrade-operations | Native rollback eligibility or blue/green failback |
| Adjust capacity after checking demand and availability | 06-scaling-strategies | Capacity and scaling policy |
| Debug failing pods | 16-troubleshooting-playbook | Symptom-based diagnosis |
Contributing
When adding new operational procedures:
- Include practical code examples (Terraform, YAML, bash)
- Provide validation steps for each procedure
- Document rollback procedures where applicable
- Add relevant PromQL/LogQL queries for monitoring
- Cross-reference related concept documentation