Low Priority Pod Guaranteed Run Provisioning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
During peak load periods of high priority workload deployments, orchestration platforms like Kubernetes often evict low priority pods, leading to disruptions and unavailability of low priority services, which can cause user dissatisfaction and service level agreement violations due to the lack of scheduling pattern analysis and resource optimization.
Innovation Solution
A computer-implemented method that identifies low priority pods for guaranteed runs by analyzing historical and resource information, informs users of potential downtimes, and provisions these pods on different host nodes or resizes them to ensure availability, using an artificial intelligence component to predict peak loads and estimate resource needs, thereby preventing evictions and maintaining service assurance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If low priority pods are evicted during peak load periods of high priority services, then resource availability for high priority services is improved, but service availability and reliability of low priority services deteriorate
Solution Approach 1:
The system performs preliminary analysis of historical workload patterns and resource utilization to predict future peak load periods before they occur. By identifying predicted peak load periods in advance, the system can proactively notify users and prepare alternative pod placements before evictions are necessary, thus maintaining service availability while ensuring resource availability during actual peak periods.
Solution Approach 2:
The system creates a copy of the low priority pod and provisions it on a different host node before the original pod is evicted. This replicated pod ensures service continuity and availability even when the original pod is removed to make resources available for high priority services during peak load periods.
2Reliability
If historical and resource information analysis is performed to identify pods for guaranteed runs, then service assurance and availability are improved, but system complexity and computational overhead increase
Solution Approach 1:
The system implements a multi-functional analysis mechanism that simultaneously evaluates historical workload patterns, current resource utilization, predicted peak load periods, and pod priority levels. This universal analysis approach consolidates multiple evaluation functions into a single integrated process, reducing overall system complexity while comprehensively assessing pods for guaranteed run designation.
Solution Approach 2:
The system automatically performs the analysis of historical and resource information, predicts peak load periods, identifies suitable low priority pods for guaranteed runs, and provisions replicated pods without requiring manual intervention. This self-service automation reduces operational complexity and computational overhead by eliminating the need for manual analysis and decision-making processes.
3Duration of action of stationary object
If low priority pods are provisioned on different host nodes before eviction, then service continuity is improved, but resource allocation efficiency and scheduling performance deteriorate
Solution Approach 1:
The system extracts only the essential function of the low priority pod (service continuity) and replicates it on a different host node. By separating the replication function from the original pod management, the system maintains service continuity while allowing the original pod to be efficiently evicted and resources reallocated to high priority services without scheduling delays.
Data Source
AI summary
Service assurance is provided. A low priority pod corresponding to a low priority service in an orchestration platform that is to be evicted due to a predicted peak load period of a high priority service is identified based on analysis of historical and resource information. The low priority service corresponding to the low priority pod that is to be evicted due to the predicted peak load period of the high priority service is marked as an assured service for a guaranteed run in response to receiving an input from a user who was notified regarding eviction of the low priority pod. The low priority pod corresponding to the low priority service that is to be evicted due to the predicted peak load period of the high priority service is provisioned on a second host node prior to the eviction of the low priority pod from a first host node.