Cloud Native Monitoring

Table of contents

1. 1. The Three Phases of Observability: An Outcomes-Focused Approach
   1. Remediating at Any Phase
   2. The Three Phases Illustrated
2. 2. Why Do You Need Metrics?
   1. Metrics as a Starting Point
   2. The Case for Metrics
      1. Metrics Provide an Efficient Snapshot of the System
      2. Metrics Are Easy to Use
      3. Metrics Let Us Aggregate Data Quickly
      4. Metrics Allow Us to Create Alerts
   3. Metric Data Is Growing in Scale
      1. Understanding Cardinality and Dimensionality
      2. Cloud Native Systems Are Flexible and Ephemeral
      3. Cloud Native Services and Systems Have Greater Interdependencies
   4. The Risk of Losing Focus on Outcomes
3. 3. The Rise of Open Source Metrics
   1. User-Controlled Metric Data Collection
   2. Prometheus: The Good and the Not-So-Good
   3. Horizontal Scalability
      1. Self-Managed Remote Storage Options
      2. Fully Managed SaaS Options
      3. Choosing Between Self-Managed and Fully Managed
4. 4. Strategies for Controlling Metric Data Growth
   1. Retention
   2. Resolution
   3. Applying Resolution and Retention in M3
   4. Aggregation
   5. Applying Aggregation in M3
   6. Conclusion
5. 5. Building Great Metrics Functions
   1. Out-of-the-Box Standard Instrumentation and Dashboarding
   2. Adding Business Context to Standardized Metrics
   3. Creating SLOs from Standardized Instrumentation
   4. Monitor the Monitor
   5. Write and Read Limits
   6. Safe Ways to Experiment and Iterate
6. Conclusion
7. About the Authors