Explore the following questions to learn more about how Sedai works with your cloud resources.
How does Sedai ensure resources under its purview are optimized for cost, performance, and reliability?
Sedai's autonomous cloud performance and optimization are driven by settings at three levels: account, group, and resource. Settings can be configured to optimize resources with a goal to decrease cost, improve performance, or both. Navigate to Settings > Topology and select Optimization to set your preferred goal.
Yes, Sedai continuously adapts to changes in application behavior. Factors like code changes or configuration adjustments can impact an application's performance. Sedai monitors these and adjusts the application model to recommend & apply the most efficient configuration per application.
Will Sedai intelligently match vertical and horizontal sizing to optimize the overall resource configuration?
Sedai continuously finds the optimal configuration for resources, including optimizing both vertical sizing (e.g., CPU, memory) and horizontal sizing (e.g., HPA, autoscaling). This allows an optimal configuration across multiple pods or tasks for your resource to respond to unexpected traffic spikes and seasonality patterns.
In addition to managing settings at the account, group, and resource level from Settings > Topology in the platform, Sedai also utilizes APIs to change settings at the group level, providing a flexible and customizable way to manage your configurations.
Can Sedai accommodate specialized requirements for container workloads, such as minimum replicas or specific CPU and memory constraints?
Yes, Sedai accommodates the specific needs of your applications by considering constraints for workloads, including minimum replicas, saturation values, and auto-scaling preferences.
Does Sedai provide recommendations for improving code to make workloads more amenable to autonomous optimization?
While Sedai does not directly read application code, it can indirectly suggest code issues through Release Intelligence. All code or configuration changes are analyzed for performance, cost, and availability deviations and reported to developers to provide insights into code-related issues. Releases are analyzed and scored against previous releases.
Can Sedai automatically detect and address scenarios where pod and thread sizing must align with input dimensions and characteristics?
Custom remediation workflow or Sedai IFTTT can be used for specialized workload requirements. Sedai can also accommodate custom performance metrics. If this behavior can be reflected using metrics, Sedai can use it to influence its recommendation.
Does Sedai limit optimization actions based on fault domains? Will application availability be affected during optimization actions?
Sedai's changes are localized to a single fault domain at a time, and even within that domain, Sedai operates safely to avoid disruptions to availability and application performance.
Sedai creates profiles specific to each application or deployment, considering traffic patterns and release cycles. Operations will only take place during safe execution windows. You can also update settings using APIs or tags to pause autonomous mode or specify execution windows.
Safety is Sedai's number one priority for all autonomous actions. The platform's AI models are based on thorough analysis of resource seasonality and dependencies, which informs the system as to whether availability would be impacted. Prior to executing an autonomous action, Sedai performs a series of safety checks and only proceeds if it can guarantee secure execution. Learn more about safety and security
Sedai sends critical notifications to your preferred notification providers in case of unusual availability issues, even when caused by external sources. The system autonomously halts actions if potential availability problems are detected. Learn more about notifications