Monitoring Distributed Systems

Calculator Example

Explain 😄

• Metrics
• Events
• Centralised Logging
• Structured Logging
• Traces
  • Not Just for Remote Calls! Trace function calls.
    • Be careful with your runtime

Task 1

Implement, in your teams, your calculator function in the language of your choice.

Task 2

Add a /_liveness endpoint that confirms if you're application is alive. Ping/Pong is fine

Task 3

Add a /metrics endpoint that exposes metrics in Prometheus exposition format

Ensure these metrics are being consumed by Telegraf and inserted into InfluxDB

Task 4

Deploy all services to Kubernetes

Task 5

Centralise your logging. You can use InfluxDB/Telegraf, or you can use ELK if you prefer

Task 6

Add "Fault Injection" into your services and publish those characteristics to InfluxDB (Events)

Task 7

Add Distributed Tracing with Jaeger