Overview

Leveraging principles from Machine Learning and Topological Behavior Analysis, the Performance Forecasting Dashboard (Not Available in SIOS iQ free edition) provides the user with a 7-day forecast of both Compute and Storage level Performance Issues that will arise due to the environment configuration. Specifically, the compute (CPU and vMemory) and storage resources configured on the Hosts and Datastores, respectively, are consistently stressed by varying workloads running on the VMs that, therefore, have the collective potential to impact performance in the current infrastructure due to resource overprovisioning. These workloads exhibit behavioral trends that are characterized via SIOS iQ’s machine learning analytics and then leveraged to predict precise time frames in which Hosts and Datastores are expected to experience Performance Issues. The Performance Forecasting Dashboard presents a daily visual overview of these predicted issues in a concise format, providing the user with a 7-day forecast at a glance. Accompanying the 7-day forecast are an Issue list and a Preview pane that together enable individual issues for a selected future date to be examined in greater detail.

Dashboard Components

The Dashboard contains the following components:

  • 7-Day Forecast – daily forecast tiles, which display:
    • The future week day and date
    • The number of issues predicted to occur on the selected date
    • A ring whose color indicates the relative number of performance issues predicted, given the current environment configuration (green, yellow, orange, or red, in increasing severity)
  • Issue List – the list of predicted Performance issues for the selected day
  • Preview Pane – a preview of the selected issue having the following tabs:
    • Details – provides a snapshot overview of the issue, including thumbnail graphs of the symptom and object topology, issue type (Workload CPU/Memory or Latency) and details, time frame(s) where the issue is predicted to occur, and recommendations to address the issue
    • Impact Analysis – shows the root cause and impacted objects (Host, Datastore, VMs) related to the predicted issue. The List View provides a summary list of the impacted object(s), while the Topology View provides an interactive graphical representation of the relationships among these objects.
    • Symptoms – provides information about the symptom(s) associated with this issue. The graph provides an overlay of the predicted time frame(s) associated with this issue.

The Compute Contention Forecast Issue

Severity Root Cause Type Layer Symptoms Impacted Objects
VM(s) Compute Contention Forecast Compute Workload (CPU/Memory) Host

What does that mean? And what should I do with this information?

This Warning issue indicates that a Host has been predicted to experience performance issues within the next 7 days. One of the following recommendations is provided:

  • Recommendations:
    • Relocate vm from host A to host B to mitigate forecast resource contention issues.
    • Add compute resources to the host associated with this issue to mitigate forecast resource contention.

The Storage Contention Forecast Issue

Severity Root Cause Type Layer Symptoms Impacted Objects
VM(s) Storage Contention Forecast Storage Latency Datastore

What does that mean? And what should I do with this information?

For each Cluster, the service level for Datastore Latency may be specified via the Storage Policy. SIOS iQ leverages this information to customize Performance expectations for related Datastores and provide relevant forecasts.

This Warning issue indicates that a Datastore has been predicted to experience performance issues within the next 7 days. One of the following recommendations is provided:

  • Recommendations:
    • Relocate virtual disk on vm from datastore A to datastore B to mitigate forecast resource contention issues.
    • Add storage resources to the datastore associated with this issue to mitigate forecast resource contention.

Whenever possible, SIOS iQ will provide precise recommendations on how to rebalance the environment in order to avoid the forecast issue. However, if a precise mitigation strategy for the forecast issue could not be found, the user is advised to add resources to the recommended Host or Datastore to relieve future resource contention.

Feedback

Was this helpful?

Yes No
You indicated this topic was not helpful to you ...
Could you please leave a comment telling us why? Thank you!
Thanks for your feedback.

Post your comment on this topic.

Post Comment