This Tech Talk shows how to build and tear down a lab environment from Google Cloud Shell so you can practice safely.
What you will learn
- Build a temporary lab environment from Google Cloud Shell
- Tear down lab resources cleanly when finished
- Practice migration tooling safely outside production
Transcript
This walkthrough builds a small Google Cloud lab, runs agentless discovery with machine stats and Tidal Tools, validates results in Tidal, then tears the lab down.
What you will stand up
The lab creates three servers inside the Google Cloud free tier and uses Google Cloud Shell as the control plane. You will install Tidal Tools for loading data into Tidal, and machine stats (an open source library) to capture CPU, RAM, and storage utilization across the fleet.
Start from the workshop repo
Open the Tidal Migrations machine stats workshop repository on GitHub and use Open in Google Cloud Shell. Trust the repository when prompted, wait for Cloud Shell to provision, and let it clone the project. The same steps work on any Linux host—not only Cloud Shell—including VMs on your network or marketplace images.
Follow the guided steps
Use the in-shell tutorial to select a dedicated project and authorize Google Cloud for the session. Install machine stats (a Python module that runs well on Ubuntu, Debian, and other Linux distributions, and also on macOS). Confirm the install with a help command.
Enable permissions to create VMs, then deploy the three lab instances. List instances and note host names and IP addresses. Generate an SSH key pair, deploy the public key to each VM, and create the inventory host file from the instance list.
Discover and sync to Tidal
Run machine stats against the host file to collect utilization metrics as JSON. This is the same pattern used to discover large server estates.
Install Tidal Tools with the one-line Linux installer, confirm the version, and run tidal login to your Tidal instance. Pipe machine stats output into tidal sync servers so lab inventory lands in Tidal for discovery and assessment.
Validate in Tidal, then clean up
In Tidal, open discovery jobs and confirm the sync completed with three servers. In assessment, find the imported instance-pool servers and review basic facts such as OS and memory.
Machine stats is open source on GitHub, which makes it easier to share the exact remote commands with security teams. For ongoing collection, use the repeat machine stats flow to poll on a daily or more frequent cadence.
Finally, destroy the lab resources so you do not leave VMs running. In the free tier this exercise should cost little or nothing. You will have proven an agentless path from lab servers to utilization metrics to Tidal inventory in a few minutes.