Kubernetes classroom Notes – 13/July/2024

Site Reliability Engineering

• Refer Here for sre books published by google.
• SRE is an engineering process on how google runs production systems
• Engineering ideas were largely adopted by customers of google and also other enterprises and now we have a job role called as SRE
• Refer Here for presentation on SRE

Observability

• Observability runs on collection 3 major informations about applications
  • metrics: A numerical value that represents some collected metric (cpu, memory, latency, error rate)
  • logs: A text record
    • levels:
      • information
      • warning
      • error
      • debug (verbosity levels)
  • traces
• We integrate the above with actionable alerting system.
• Centralized log aggregation tools:
  • Elastic Search (logstash and beats)
  • Splunk
  • Fluentd
  • datadog (sass product)
• metrics:
  • New Relic
  • Metric beats => Elastic Search
  • nagios & zabbix
• Tracing (APM)
  • app dynamics
  • elastic search apm
• How to acheive observability
  • Fluentd
  • Prometheus
  • Grafana