Infrastructure

Edge Node Scaling

The Spendplane Node Scaling engine manages the compute resources of your Spendplane Edge Nodes. It ensures high availability for your AI agents while minimizing idle costs.

How do I route v0 traffic to Llama-3?

Automatic Vertical Scaling

Spendplane dynamically adjusts the resource limits of your gateway workers based on memory usage and request latency.

STATUS: OPTIMIZED RES-LOAD

Horizontal Node Clusters

For enterprise deployments, the Spendplane Cluster distributes gateway workers across multiple Spendplane Edge Nodes to provide region-wide high availability.

DISTRIBUTION: BALANCED (3 NODES)

Configuration (Auto-Scaling)

You can define your node scaling policy in your global spendplane.config.ts. This allows you to set hard limits for cost control while enabling bursts for heavy LLM processing.

Spendplane Edge Node Config
# spendplane.config.ts
{ "scaling": { "min_nodes": 1, "max_nodes": 10, "target_cpu_utilization": 75, "burst_support": true } }
Performance Optimization

Scaling events warm standby workers proactively when Spendplane Gateway detects an uptick in prompt frequency.

Trace Hub: Manual v1.0.4 / Compute Eng Verification: SCALED