Optimize pod resources and applications, together
Balance cost targets and performance optimization goals
Akamas is designed to achieve complex user-defined goals. Users can set resource usage, application performance, and response time goals, while respecting latency, error rate and SLO constraints.
Prevent dangerous application configurations
Akamas AI learns continuously from system and application signals, such as response time and error rates, finding stack configuration sets that prevent downstream issues such as out-of-memory errors.