Analytics Portal - csap-platform/csap-core GitHub Wiki

The CSAP analytics portal retrieves data from the CSAP Events Service to provide extensive operational and performance reports. Everything from the aggregate CPU consumed across a specified application, host resource (cpu/disk/memory,..), service resources and work produced, to user activity reports.

Contents:

click diagrams to enlarge to enlarge

Aggregation Analytics

Screenshot Description

Landing Page

  • aggregate reports (cpu, user activity, log rotations, ...) across all services on all hosts
  • ability to compare different application lifecycles - how testing in non prod relative to production loads.

Host Summary Report

  • table view allows sorting by attibute to identify which host is consuming the most, the least, etc
  • configurable time periods, filters, etc.

Host Summary Compare

  • selectable time periods allow for before / after comparisons
  • improvements and worsening are quickly identifiable

Host Attribute Comparison: CPU Load

  • Any of the attributes collected can be compared to each other, or previously collected
  • shown is cpu load

Host Attribute Aggregate Trending: CPU Load

  • Addresses: How is my application Host CPU Load trending over X weeks.

Service OS Resources - All Services, All Instances

  • Similar to Host performance, OS resources for services can be viewed and analyzed

Service OS Resources - 1 Service, All Instances

  • across a wide number of instances - there can be significant variability in resources consumed
  • highlights a single service deployed on many hosts

Service Threads

  • Showing relative thread count across all services

Java Summary - All services

  • For java based services, key fields (jvm heap, user sessions, etc.) are also available.

Java User Sessions - session leak

  • Note the active user growth.
  • It was only occurring on specific host instances

click diagrams to enlarge to enlarge

Time Line Analytics

Screenshot Description

Host Performance Comparison

  • each line on the graph reflects resource consumption per host
  • NOTE: graphs are shown stacked; but checking lines will unstack

Smart Select

  • when dealing with large numbers of hosts - using smart select to filter outlier results.
  • shown: top 10 hosts with highest load average compared the 2 hosts with the lowest load average

Service - OS Performance

  • multiple instances of a given service are deployed for scalability / availability
  • each line on the graph reflects resource consumption for a specific service instance
  • shown: the same service - but deployed on different hosts
  • based on traffic distribution, specific functional use cases - resource profiles over time can/will vary
  • NOTE: graphs are shown stacked; but checking lines will unstack

Service - Java Performance

  • Service shown: csap data service (receive data from CSAP agents)
  • reference: CSAP Events Service
  • dropdown filters available: http data, jvm heap, java thread. jvmHeap, unstacked:

Service - Application Performance

  • Service shown: csap data service (receive data from CSAP agents)
  • reference: CSAP Events Service
  • Note: application data is domain specific

click diagrams to enlarge to enlarge

Events Browser

Screenshot Description

Launch Points

  • Csap Events dashboard is integrated throughout the [Management Portals]
  • shown: the [Application Portal] landing page: clicking on event count or user activity opens the event browser with filters preselected
  • reference: CSAP Events Service

Browser: Filtered by User Events

  • Events browser shows events received
  • shown: events filtered by /csap/ui

Browser: Filtered by host reports

  • Launched from Analytics Landing Page
  • shown: Host Summary table - filterd by cores >= 4, followed by reports
⚠️ **GitHub.com Fallback** ⚠️