Platform

Key Platform Features

Everything OptOps does to cut your Kubernetes bill autonomously — and the safety mechanisms that let you trust it.

move_up

Container Live Migration

Move running pods between nodes without dropping a connection — memory and process state preserved via CRIU, with the same pod IP and unbroken TCP sessions on AWS.

Learn more arrow_forward

bolt

State-Preserving Spot Rescue

When a spot node is reclaimed, workloads are live-migrated off it before the deadline — proven against real AWS FIS interruption drills.

balance

Autonomous Consolidation

Bin-packing plans that respect PDBs, affinity, and volumes — gated so a plan only executes if it actually saves money.

content_cut

Continuous Rightsizing

Requests that track real usage, continuously, with OOM spike protection that scales workloads back up before kills bite.

savings

Spot Automation & Resilience

Diversified spot pools, automatic on-demand fallback, and a stockout circuit breaker for capacity crunches.

cloud

Karpenter-Grade GKE Engine

The same autoscaling engine on AWS and Google Cloud — including the GCP provider upstream Karpenter never shipped. Stock-Karpenter compatible, no lock-in.

payments

Real-Time FinOps Visibility

Track cost by cluster, namespace, workload, node type, and instance class — no tagging required. Multi-cluster from a single dashboard.

bedtime

Sleep Schedules

Dev and staging clusters that sleep when your team does — cron-based scale-down with one-click schedules.

visibility

Shadow Mode & Graduated Autonomy

Start read-only, watch every planned action in shadow mode, then enable automation per capability and per cluster.

receipt_long

Honest Savings Reporting

Realized savings and remaining opportunity reported separately — never netted, never inflated, never double-counted.

tune

Console-Tunable In-Cluster Engine

Tune the optimizer from the dashboard while every decision is made inside your cluster — not on our SaaS.

shield_lock

Enterprise-Grade Security

Read-only by default, no secrets collected, full auditability. Independent third-party security assessment completed; ISO certification in progress.

How OptOps works, end to end

check_circleInstall the full stack with one Helm command — agent, node engine, spot handler, and live migration, all default-off.
check_circleCollect read-only metrics and detect pod and node waste against real usage.
check_circleWatch every planned action in shadow mode before anything is enabled.
check_circleTurn on autonomous rightsizing, consolidation, and spot — capability by capability, with audit visibility throughout.