Introduction to Metrics Based Monitoring

During the course we acquire the entire life cycle of metric data, from the configuration of the collectors and the collection of data on the measured systems through the transport layer to the transformation (various aggregations or metadata modifications) when stored in the time series database.

We will also focus on the evaluation of the metrics and, if the input conditions are met, their subsequent processing by means of a communication interface or automatic processing by the orchestration tool. We learn the formatting of queries for different kinds of databases, their use in defining visualizations and notifications.

We will show you examples of setting data collection on real infrastructure and show you how to access the stored data, which will provide you with a good basis for further use of modern, metric-based monitoring systems.

Audience

• DevOps engineers
• IT administrators
• IT Architects

Goals

Outline

Configuration and autodiscovery of measured services

• Cloud services
• Containers
• Network devices
• Application observability

Types and kinds of metric data

• Counters
• Timers
• Histograms
• Summaries

Mechanisms of data collection and transmission

• Push vs pull mechanism
• High availability of collection

Options for storing time series

• Aggregation
• Retention
• Classification

Examples of metric collection configuration

• Collectd
• Telegraph
• Prometheus exportery

Transformations and queries on time series

• Query formats
• Expanding features
• Connection to expert systems

Alarm definition and event management

• Threshold values, hysteresis
• Root causes

Visualization of metric data

• Vectors - values at a given moment
• Matrix - the entire time series
• Grafana dashboards
• JSONNET dashboard mixes

Metric data processing

• Alarmmanager rules
• Sensu notification
• JSONNET alarm mix

Examples of queries for time series databases

• Prometheus
• Graphite
• InfluxDb

Prerequisites

• Fundamentals of Linux / Windows platform management
• Experience with public clouds and services

Technical requirements (BYOD)

• A computer with any OS (Linux, Windows, OS X)
• SSH client
• Web browser