Skip to content

Introduction to OptiPod

OptiPod is an open-source Kubernetes operator that helps you optimize resource requests and limits for your workloads. It analyzes actual resource usage and provides recommendations to right-size your containers, reducing waste and improving cluster efficiency.

Kubernetes requires you to specify resource requests and limits for your containers, but getting these values right is challenging:

  • Overprovisioning wastes cluster resources and increases costs
  • Underprovisioning causes performance issues and OOM kills
  • Manual tuning is time-consuming and error-prone
  • VPA (Vertical Pod Autoscaler) can conflict with GitOps workflows

OptiPod solves these problems by providing safe, explainable, and GitOps-friendly resource optimization.

FeatureOptiPodVPA
GitOps-safe✅ Yes❌ Webhook conflicts
Explainable recommendations✅ Yes❌ Opaque
Policy-driven control✅ Fine-grained⚠️ Limited
Recommend-only mode✅ Yes✅ Yes
Auto-apply mode✅ With safeguards✅ Yes
Update strategy options✅ Webhook or SSA⚠️ Webhook only
Memory safety guardrails✅ Yes⚠️ Basic
  1. Deploy OptiPod operator in your cluster
  2. Create OptimizationPolicy CRDs to define what to optimize
  3. Label workloads to opt-in to optimization
  4. Review recommendations stored as annotations on your workloads
  5. Apply changes manually or let OptiPod auto-apply with safety guardrails

OptiPod supports three operational modes:

  • OptiPod analyzes workloads and stores recommendations as annotations
  • You review and apply changes manually (via GitOps or kubectl)
  • Best for: Getting started, testing, GitOps-strict environments
  • OptiPod automatically applies recommendations with safety guardrails
  • Conservative defaults prevent dangerous changes
  • Change-rate limits prevent rapid oscillation
  • Best for: Production environments with confidence in OptiPod
  • OptiPod stops analyzing and updating workloads
  • Existing recommendations remain but aren’t updated
  • Best for: Temporary pauses or troubleshooting

OptiPod is designed with safety as a top priority:

  • Conservative defaults: Never reduce memory below safe thresholds
  • Policy-driven controls: You define bounds and change limits
  • Memory safeguards: Prevents OOM kills from aggressive reductions
  • Configurable safety factors: Apply multipliers to recommendations

Ready to get started? Follow our installation guide:

Install OptiPod →

Or learn more about core concepts:

Core Concepts →