Capacity Planner Release Notes 95


Overview

Released: May 2018

This release provides a number of updates allowing users to further make use of the ‘Models’ tab to configure the view of the environment, with visibility into summaries of subsets of an entire model. A workload summary dialog has been introduced that will summarise all workloads from a chosen segment of the sunburst, providing information on the sum and average of each metric associated with any workloads at that point in the hierarchy.

Enhancements have also been made to forward thinking scenario modelling, allowing the user to selectively remove individual metrics from a migration and also set CPU, memory or storage demand on a workload directly.

Highlights

Viewing different models

The models tab has been enhanced to allow the user to select from multiple available models. A model is a summary of a given time frame. For example, a summary of a particular week, day or month, or perhaps a summary of working hours only.

In the default baseline, each metric has a summary using all data within the scope of that baseline. For example, if a default baseline runs from 1st January to the 1st April, each metric will be summarized using all data in that time period. A summary is a collection of statistical measures and percentiles based on that data; 5th, 25th, median, mean, 75th, 95th, 99th, max. When looking at the CPU demand for a workload in a default baseline, the summary shown is the 95th %ile of demand over the baseline time period (assuming the user has not selected an alternative measure).

Release 94 introduced the capability to the user to change the percentile shown for each metric when viewing the default baseline. Release 95 further introduces the facility for the user to change the time period used to create that statistical summary. These different time periods are called ‘models’.

These models are created automatically on a daily basis, by default, the user can select the following models:

Model name Description
Baseline summary A summary of all data in the baseline.
Monthly summary A summary of all data foe a calendar month. The user will be prompted to select a particular month.
Daily summary A summary of all data for a single day. The user will be prompted to select a particular day.

Further models can be configured and created automatically on a daily basis. To do this, please contact your Capacity Planner representative who can configure this.

Once a model has been selected, the user can also select the measure to be used. For example, selecting a daily model of 19th April 2018 and then a CPU measure of max will configure the sunburst to calculate available headroom using the peak CPU of every workload and host/server on the 19th April.

The following screenshots demonstrates typical contents of the Model tab.

When a model is selected, all timeseries and box-plot representations will be updated to indicate what time range is being included in the model. The following example is a drill-through launched from a model that is configured to show the summary of calendar month February 2018. The segment highlighted on the lower part of the dialog shows those dates included in the model.

This highlighting is also shown on timeseries data. The following example is a CPU timeseries for a host where the model is configured to show the summary of calendar month February 2018.

Please note, the changing of models and summary measures is for baseline reporting and data exploration purposes only. Forward thinking scenarios use a mechanism trained on the default metric measures. Should the user attempt to create a new model using a baseline with non-default metric measures, those measures will revert to the defaults when the scenario is created.

Workload summary

On the right click menu, under the ‘workloads’ sub-menu, a new ‘Workload summary’ option is available. This will present a summary of all metrics on all workloads from the segment that was selected in the hierarchy. It will present each metric, the number of VMs from that point of the hierarchy downwards, the measure being used for that metric in the current model, the number of workloads that have that metric and then the sum and average for that metric. The following is an example where a small subset of workloads in an estate have specific application metrics configured.

Excluding metrics from migrations

The user can now exclude metrics from any scenario operation that involves the movement of workloads (for example, cluster redistribution or migration). This is particularly useful when memory.consumed.average is used in a VMWare environment. This metric can often over-state the use of memory for particular applications and the user may wish to optimise a cluster estate while ignoring the aggregate consumed memory. The following in an example of a redistribution operation using minimum number of hosts within a cluster and ignoring the memory.consumed.average metric.

Setting demand

Until now, users have had the facilty to ‘add demand’ to workloads that will effectively ‘inflate’ the demand of a metric by a chosen absolute value or percentage. This release introduces the facility to specifically set demand for a workload. This can be particularly useful when modelling a greenfield site where hardware configuration and average VM configuration is known, but no demand profile information is available. Another area where this can be useful is in the modelling of scenario when all provisioned VMs use the resources they have been provisioned, and demonstrating that capacity in a cluster may not be sufficient to support this demand.

This menu item is available in the scenario modelling right click menu and will apply to all workloads. It is currently restricted to CPU, Memory and shared and local storage. The following is an example of setting all metrics of VMs to 50%.

Further enhancements

The following enhancements have been introduced in this release for usability purposes:

  • The user can now delete application demand templates.
  • All time series for workloads, hosts and servers will now show capacity where relevant.

Other releases

Click the links below to view other versions of Capacity Planner release notes.

Capacity Planner Release Notes 100 Released: May 2019
Capacity Planner Release Notes 99.2 Released: March 2019
Capacity Planner Release Notes 99 Released: December 2018
Capacity Planner Release Notes 95 Released: May 2018
Capacity Planner Release Notes 94 Released: March 2018
Capacity Planner Release Notes 93 Released: January 2018
Capacity Planner Release Notes 92 Released: December 2017
Capacity Planner Release Notes 91 Released: October 2017
Capacity Planner Release Notes 90 Released: September 2017