Wave Certified

Per-workload ML pod scaling that replaces HPA's static math formula. Each deployment gets its own model that learns its traffic and resource patterns — delivering 2x faster scaling decisions than HPA without manual threshold tuning.

Features:

• Autopilot — Per-workload ML scaling, more accurate than HPA's CPU/memory thresholds
• Autopilot Scheduler — Time-based pre-scaling for planned events via calendar UI
• Min/Max Recommendation — Data-driven HPA replica count optimization

AI-driven resource optimization that analyzes actual container usage to eliminate over-provisioning. Recommends optimal CPU/memory requests and limits, reducing wasted resources by up to 40% while ensuring application stability.
 

Features:

• Smart Sizing — Automatic resource recommendations with In-Place Pod Resize support; pairs with Autopilot instead of conflicting with it

WASM-based priority traffic shaping on Istio that protects critical user-facing traffic during overload. Workload-aware queueing keeps checkout, payments, and login flowing while bulk traffic gracefully degrades — no manual Envoy filters required.

Features:

• Wave Flow — Per-route priority shaping across Istio Sidecar, Ambient, and Envoy Standalone
• NetFunnel Integration — Workload-aware virtual waiting room that activates automatically during surges

ML-powered analysis of cluster and workload telemetry that surfaces issues before they page someone. Forecasts capacity, predicts memory leaks, and identifies cost-cutting opportunities — all without writing custom Prometheus rules.
 

Features:

• Cluster Resource Forecast — 7–30 day cluster capacity forecasts
• Memory Leak Detection — Predicts OOMKill events hours-to-days in advance
• Pod Scheduling Delay Detection — Per-workload latency breakdowns
• CPU Utilization Analysis — Workload-level CPU breakdowns with label segmentation
• Idle Node Detection — Surfaces underutilized nodes for cost reduction

Operational completeness layer covering the rest of Day 2: persistent volume lifecycle, configurable alerting, and programmable automation hooks for Platform Engineering teams.
 

Features:

• PV Auto-Expansion + Capacity Forecast — Expands PersistentVolumes before they fill; forecasts exhaustion 7–30 days out
• PV Cleanup + Unused PV Detection — Reclaims storage costs from orphaned and idle volumes
• Alerts — Configurable webhook alerts for scaling events and anomalies (Slack, email, custom endpoints)
• Programmable API — RESTful API for everything Wave does, for pipeline integration