Management Tools

Better monitoring and logging for Compute Engine VMs

#compute

Over the past several months we’ve been focused on improving observability and operations workflows for Compute Engine. Today, we are excited to share the first wave of these enhancements are now available. These include:

  1. Significantly improved operating system support for the Cloud Monitoring and Cloud Logging agents.

  2. The ability to rapidly deploy, update, and remove agents to groups of VMs, or all of your VMs, by policy, with as little as a single gcloud command.

  3. New VM-specific features within the Cloud Monitoring console, which we’ll discuss in an upcoming blog post.

Understanding agents

Agents remain a key way to get fine-grained visibility into a virtual machine’s host operating system, and applications running on Compute Engine are no different. 

Out of the box, every Compute Engine instance (or managed instance group) provides some level of telemetry, including metrics for CPU utilization, uptime, disk throughput and operations, and networking operations. To capture more advanced operating system metrics like memory consumption and disk utilization, metrics from commonly used applications (databases, web proxies, etc.), and logs from your applications, you need to install the Cloud Monitoring and Cloud Logging agents onto each VM.

Automatic agent installation and management

Because agents are so essential in VM environments, we’ve automated the process of installing, updating, and removing the Cloud Monitoring Logging agents onto groups of Compute Engine VMs, or your entire fleet, via a new set of gcloud commands. With as little as one command, you can create a policy that governs existing and new VMs, ensuring proper installation and optional auto-upgrade of both agents. This is a great way to start using Cloud Monitoring or Cloud Logging right away, and to scale metrics and logs collection from a single VM to all VMs in a project.

These policies can be applied to Linux virtual machines now as a part of the public alpha and will apply to Windows VMs soon.

Improved operating system support

Over the past year, we’ve added Cloud Monitoring and Logging agent support to a host of new operating systems.

Linux

The Monitoring and Logging agents are now compatible with 30 of Compute Engine’s available Linux images, including:

  • CentOS 7+

  • Red Hat Enterprise Linux 7+

  • Debian 9+

  • SUSE Linux Enterprise Server 12+

  • Ubuntu 16+

With these additions, the Cloud Monitoring Linux agents can be used on every Compute Engine host operating system other than the Container Optimized OS, which has monitoring and logging capabilities built in to the OS itself.

Windows

Cloud Monitoring has been able to capture system and SQL Server metrics for Windows virtual machines since before 2015, thanks to its Windows agent. We’re currently improving the compatibility, quality, and functionality of our Windows support with a new agent that provides the following enhancements:

  • Capturing the same advanced OS metrics as the Cloud Monitoring Linux agent, rather than a smaller incompatible set

  • Compatibility with more Windows versions

  • Capturing application metrics from IIS, SQL Server, and Windows performance counters

The new agent is in preview. Please contact your account manager if you would like to participate in early tests.

Wrapping up

We hope you enjoy these improvements to Cloud Monitoring and Cloud Logging, and we look forward to bringing even more capabilities to the platform. To check out these new features, go to our documentation for these new features or to the Cloud Monitoring and Logging in the Google Cloud Console.