Advanced Signals API
Advanced real-time optimization signals for enterprise infrastructure. Combine carbon intensity, spot pricing, and electricity costs into intelligent scheduling decisions.
What Is an Optimization Signal?
A signal is any time-varying input that makes one moment or one location a better place to run a workload than another. Most infrastructure schedules as if every hour and every region were identical. They are not. The carbon intensity of a grid can swing by a factor of five or more over a single day, spot capacity is priced by supply and demand and commercial electricity tariffs change on a published schedule.
The Advanced Signals API resolves these inputs into a single decision surface, so a scheduler can answer one question: given everything known right now, when and where should this job run?
Carbon intensity
Grams of CO2 per kilowatt-hour on the grid serving a given region, plus a forecast of where it is heading.
Price signals
Spot instance pricing and time-of-use electricity rates, which frequently move in the same direction as carbon but not always.
Workload constraints
Deadlines, data residency and hardware requirements that bound which of the theoretically better windows are actually reachable.
Average or Marginal Intensity?
Carbon intensity comes in two flavors and the difference changes what your scheduler should do. Average intensity describes the mix of everything currently generating on a grid. Marginal intensity describes what generator actually responds when you add one more unit of demand. A grid can look clean on average while its marginal generator is a gas peaker.
Neither number is wrong. They answer different questions, and which one belongs in a scheduling decision depends on whether you are reporting a footprint or trying to change one. We work through the trade-off in detail in Average or Marginal? Navigating Intensity Signals in Carbon-Aware Computing.
Advanced Features
Real-Time Signals
- Carbon intensity forecasting
- Spot instance price tracking
- Electricity cost monitoring
- Custom signal weighting
Intelligent Optimization
- ML-powered predictions
- Adaptive scheduling
- Multi-factor optimization
- Custom constraints
Advanced Analytics
- Detailed metrics
- Custom reporting
- Trend analysis
- ROI calculations
From Signal to Scheduling Decision
1. Ingest
Carbon intensity, spot pricing and electricity rates are collected per region and normalized onto a common time base.
2. Forecast
Point-in-time readings tell you about now, which is rarely when a deferrable job needs to run. Forecasting turns a current reading into a usable scheduling window.
3. Weight
Carbon and cost do not always point the same way. Custom signal weighting lets you declare the trade-off you want rather than accepting one we picked for you.
4. Schedule
The scheduler holds deferrable pods until conditions clear your threshold, or places them in a cleaner region when the workload permits it. Application code is untouched.
Integration Options
Infrastructure Support
- Kubernetes Integration
Native support for k8s scheduling
- Cloud Providers
AWS, GCP, Azure support
- On-Premise
Direct datacenter integration
Security & Reliability
- Enterprise Security
SOC 2 compliance & encryption
- High Availability
99.99% uptime SLA
- Data Privacy
GDPR & CCPA compliant
Flexible Pricing
Our enterprise pricing is based on your infrastructure size and optimization needs. Early access partners receive preferential pricing and direct input into our product roadmap.
Usage-Based
Pay only for what you use
Volume Discounts
Scale efficiently with your needs
Custom Plans
Tailored to your requirements
Learn More
Start with our free open-source scheduler to experience carbon-aware computing, then upgrade to the Advanced Signals API when you need enterprise-grade optimization across multiple regions and cost factors.
Need help implementing? Our consulting team can guide you through assessment, deployment, and optimization.
Further reading
- The Grid Isn't a Cluster on why software mental models mislead when applied to electricity.
- Carbon-aware regional routing for LoRA fine-tuning covering 21 training jobs placed by grid conditions.
- Frequently asked questions on integration, pricing and workload fit.
Join Our Early Access Program
Be among the first to leverage our advanced API capabilities. Early access partners receive dedicated support, implementation assistance, and the ability to shape our roadmap.