Resiliency

Strategies to increase resiliency:

• Application development practices:
  • MicroServices fault tolerance
  • Timeout non-critical functions and gracefully degrade subsets of the HTTP response
• Increase caching to reduce workload including:
  • HTTP response caching
  • Servlet caching
  • Content delivery networks (CDNs)
  • Caching products
• Proactive monitoring of problems:
  • tWAS: Hung thread detection
  • Liberty: requestTiming
• Optimize existing applications with a sampling profiler
• Tune timeouts to the maximum acceptable user response time for all transactions plus 20%:
  • IHS: ServerIOTimeout and per-URL timeouts with websphere-serveriotimeout
• React to excessive response times:
  • tWAS on z/OS: zWLM service classes
  • tWAS on other OSes: IHS with Intelligent Management for Web servers and excessive response time health policies
• Auto-scaling:
  • Cloud/OpenShift
  • tWAS on z/OS: Minimum and maximum servants
  • tWAS on other OSes: Dynamic cluster autoscaling using Intelligent Management and the Java On Demand Router
  • Liberty: Autoscaling with Collectives and Intelligent Management for Web Servers

Previous Section (Log Analysis) | Next Section (Major Tools) | Back to Table of Contents