Cloud Deployment Patterns

You are here:

This section describes how to realize industry recognized cloud deployment patterns with Sonar. Each pattern name is linked to Arcitura pattern catalog.

Rich Container Pattern

Realization

Configure Sonar monitoring agent as Windows service when building Docker container image.

Benefits

  • Use of Microsoft Server Core and Nano Server containers are supported.
  • Ability to expose container metrics to target monitoring systems on-premise and in the cloud without loosing autonomy.
  • Collecting wide range of metrics (performance counters, WMI, Windows logs) to multiple monitoring systems using single agent.
  • Push and pull model for metric collection.

Liabilities

  • When Sonar metrics configuration is included in container itself, further changes to metrics post- deployment may increase complexity. For example, loading Sonar configuration from external volume will require additional deployment automation.
  • Modification of Docker file is required to configure Sonar when building application container image.

Ambassador Pattern

Realization

Configure Sonar monitoring agent as Windows service when building Docker container image.

Benefits

  • Sonar deployments as Windows Server Core, Nano Server and Linux container are supported.
  • Ability to expose container metrics to target monitoring systems on-premise and in the cloud without loosing autonomy.
  • Exposing wide range of metrics (performance counters, WMI, Windows logs) to multiple monitoring systems using single agent.
  • Push and pull model for metric collection.

Liabilities

  • Further changes to metrics post- deployment may increase complexity. For example, loading Sonar configuration from external volume will require additional deployment automation.

Sidecar Pattern

Realization

Configure Sonar monitoring agent as Docker container with metrics to expose from another container instance.

Benefits

  • Sonar sidecar deployments as Windows or Linux containers are supported.
  • Ability to expose container metrics to target monitoring systems on-premise and in the cloud without modifying application containers and loosing autonomy.
  • Exposing wide range of metrics (performance counters, WMI, Windows logs, databases) to multiple monitoring systems using single agent.
  • Push and pull model for metric collection.
  • Ability to downsample application metrics for exposing them to container orchestrator for monitoring and autoscaling.

Liabilities

  • Further changes to metrics post- deployment may increase complexity. For example, loading Sonar configuration from external volume will require additional deployment automation.
  • Akumuli time series database and Sonar monitoring agent are required for sidecar deployment for collecting and downsampling metrics.
  • Collecting application and OS logs by sidecar is not supported.

Last Updated On March 27, 2019