Monitoring

• Consider the following application architecture
  
• The above architecture is a sample architecture to make a movie ticketing application work.
• What can go wrong here?
  • network failures
  • power failures
  • OS failures
    • Disks getting filled up
    • New OS Updates/Patches crashing
  • Hardware issues
  • Services crashing
    • code issues
    • High CPU utilizations
    • High Memory
    • High I/O on disk
  • Performance issues
  • Saturation:
    
  • Latency: Time Taken by packet to reach the server (journey time)
  • High CPU and Memory, Disk Issues
• When things go wrong what can be done for quick resolutions ?
  • We cant avoid all failures, but we need quick resolutions
• Reducing Problems in Applications might not be possible. But on a broader note there are two approaches
  • Preventive approaches
  • Reactive approaches
• To acheive the faster resolution we need to monitor. We are supposed to monitor
  • Server/Infrastructure
    • Basic health check for server/infra up or not
    • Some metrics to collect
      • CPU Utilization
      • Memory Utilzation
      • Free disk space
  • Application Monitoring
    • Application home page accesible or not
    • logs of applications