Managing the capacity of your GCP infrastructure



What can you do with TrueSight Capacity Optimization?

TSCO_goals.png

As a Capacity Planner or GCP Technology Specialist, you can use TrueSight Capacity Optimization to configure, administer, and manage the capacity of your GCP infrastructure. 

gcp_use_case.png

As the flow diagram illustrates, the data source (Google Cloud Platform - GCP API ETL) collects data from the GCP console. The collected data is transferred to the Capacity Optimization data warehouse where it is processed, and then displayed on the user interface. You can use the product features to review, analyze, and manage the capacity of your GCP infrastructure. 

The following sections describe how you can achieve these goals:

Managing the capacity of GCP infrastructure

You can analyze and manage the capacity of your GCP infrastructure elements by using the GCP VM Instances view. For the GCP data to be available in the view, the Administrator must first set up the data sources to collect data. 

Step 1. Collect data and install the views

As an Administrator, you can use the Google-Cloud-Platform-GCP-API-Extractor to collect data from the GCP VM instances.

  • Configure and run this ETL to collect the required configuration and performance metrics from the VM instances.
  • In addition to the required metrics collected by using the GCP API Extractor, you can create a Stackdriver account and install the Monitoring agent on each of your virtual machines to collect additional performance metrics from them. These metrics are useful for investigating the capacity-related issues that might occur in your Google Cloud environment. When you run the GCP API ETL, these metrics are imported into the TrueSight Capacity Optimization database. For more information, see Collecting-additional-metrics-using-the-Stackdriver-monitoring.

(Optional) You can also use the BMC-TrueSight-Capacity-Optimization-Gateway-VIS-files-parser to collect additional metrics, more accurate memory utilization metrics, and performance metrics at a higher granularity. Before you use this ETL, you must instrument the VMs. 

Instrumenting VMs

The memory utilization value that is collected from an instrumented VM is based on the actual memory of the VM. The Agent collects resource consumption breakdown at process or workload level and helps you to detect specific in-guest OS level resource constraints (for example, in-guest paging due to the physical memory configuration of the VM being too low).

BMC recommends that you instrument your business-critical VMs to collect OS-level memory usage values. 

To instrument a VM:

  1. Install a Capacity Agent inside the VM from which you want to collect metrics
  2. Configure the Gateway Server and Capacity Agent to initiate data collection. For more information, see Collecting-data-via-Capacity-Agents
  3. Configure and use the out-of-the-box BMC-TrueSight-Capacity-Optimization-Gateway-VIS-files-parser to collect the required metrics from the VM.

After data collection starts, data is loaded in the data warehouse daily and Indicators are available in the Workspace.

As an Administrator, you must install the GCP views and Business Services view and grant the necessary permissions to Capacity Planners and GCP Technology Specialists to access these views. While creating the service pools, ensure to select GCP domains to view GCP only data.

Step 2. Analyze the collected data

To get a high-level view of the GCP infrastructure usage and health at the business services and service pools level, use the Business Services view.

For detailed analysis, use the GCP-views.

The following common use cases are described here:


Understand the usage and health of your GCP VM instances

Review and analyze the VM-Instance-page-in-the-GCP-Data-Explorer-view to determine the usage and health of your VM instances.

From the VM Instance filter, select the instance that you want to view data for detailed analysis.

Determine and analyze the available resources and their utilization per VM instance

Review and analyze the relevant metrics on the VM-Instances-page-in-the-GCP-VM-Instances-view to determine the available resources and their utilization per instance. For example, utilization metrics for CPU, Memory, and network.

Performing advanced analysis

The earlier sections explained how you can use the out-of-the-box capacity views to manage your environment. These capacity views help you analyze your GCP infrastructure using a predefined set of metrics.

To perform advanced analysis on the imported GCP data, such as identifying specific performance issues, trends, and bottlenecks, you can use Analysis.

About Analysis

An analysis is a visual tool that you can use to identify the behavior of a set of metrics and the relationships among them. Each analysis can focus on the business driver metrics of an application, on the performance of an application's systems, and on the events related to an application. Analyses can also be used to compare performance and business driver metrics to determine a system's behavior under load. 

Here are some use cases for which you can create and use Analyses:

For more examples, see Creating-an-analysis.

Analyze the disk activity of a GCP VM over time and understand the trend

Analyse the resource utilization pattern of GCP VM instances

  1. Create an analysis. 
    See sample configuration values:
    gcp_vm_resource_utilization_01.png
    gcp_vm_resource_utilization_02.png

  2. Review the analysis results.
    The analysis results are shown in a tabular format. The summary table shows the resource utilization metric values (disk and network) for the virtual machine. The page shows a table for each virtual machine.
    gcp_vm_resource_utilization_03.png
    gcp_vm_resource_utilization_04.png


Managing future demand

By using the capacity views and analysis charts, you can analyze the data of your existing capacity. To predict and plan your IT resource needs, use Models.

Predicting the behavior of your resources

Use Models to predict service performance and obtain forecasts of historical series of metrics, including deep details on the modeling techniques used in forecasts and how to interpret the results of model runs.

About models

A model is a simplified mathematical description of service components that evaluates historical data, predicts future behavior, and simulates what-if scenarios. Models are always built on existing data and analysis. After you create a model, define scenarios to perform multiple predictions under different conditions. With Models you can forecast and model changes in service demands.

For more information, see Modeling-capacity-usage.

How to determine when a resource (CPU Utilization) of a Google Cloud VM instance reaches threshold


 

Tip: For faster searching, add an asterisk to the end of your partial query. Example: cert*