How to build two of the most common hybrid cloud monitoring dashboards in Stackdriver

One of the most powerful features of Google Stackdriver is it’s visualization capabilities. It is certainly one of the leaders in next-gen visualization and brings the monitoring industry one step closer to observability. So why not use Stackdriver’s superpower on all of the things in your environment? From applications like SAP to databases like Microsoft SQL or on-prem servers, storage, networking and load balancers–Stackdriver can help you trend and troubleshoot it all.

But where should you get started? This blog will walk you through how to build two of our most commonly used Stackdriver dashboards for yourself.

Step 1: Get all your operations data into Stackdriver
The first step to dashboarding is data collection. There are a few ways to go about doing this, but the easiest way to approach this task for Stackdriver is to use Blue Medora BindPlane. This monitoring-integration-as-a-service is free to Google Stackdriver users. Once configured it handles all of your operations data management for nearly 150 different sources through a single, self-maintaining connection. You’ll never have to install and update a bunch of agents or plugins.

To configure your free BindPlane account, visit the Google Cloud Platform Marketplace.

Step 2: Build a single-environment dashboard
After you’ve started to bring custom metrics into Stackdriver using BindPlane, you still have to create dashboard content that gives these metrics purpose. One of the best ways to use your new dataset is to create a full-stack dashboard for one or more of your critical applications that are still running on-prem. To get started building your first dashboard in Google Stackdriver:

  1. Click Dashboard from the left-side navigation
  2. Select Create Dashboard
  3. Give your dashboard a name
  4. Click Add Chart
  5. Select the first resource type and metric you wish to track from the available list
  6. Name the chart
  7. Click Save
  8. Repeat steps 4–7 for all your application stack KPIs

In Figure 1, I’ve built a full-stack dashboard that includes a Microsoft SQL Server database hosted on VMware vCenter and using Pure Storage. Obviously, your stack will have different types of specific resources, but this list of recommended KPIs is a good baseline to get started in many stacks.

Database metrics
Active connections
Disk wait time
Disk operations

Host metrics
Average Read/Write Latency
CPU Usage
I/O Rate
Memory (Bytes)

Storage volume metrics
Average Read/Write Latency
I/O Rate

Monitoring On-Premises Infrastructure w/ Stackdriver

The above dashboard was built out to represent full visibility into the stack. You should be able to drill down through the stack to understand how our workload is affecting our vCenter host and our storage array. All of this uses the custom metrics being sent from BindPlane into Stackdriver.

Want to learn how to build a multi-environment dashboard? Keep reading…