Gremlin and GCP Batch (VM-backed jobs)

Test GCP Batch reliability with Gremlin. Simulate process crashes and VM failures to validate job rescheduling.

Google Cloud Batch provisions VMs, runs your work, and tears them down. That efficiency comes with a diagnostic problem: when a task doesn't complete, the instance running it may no longer exist by the time anyone investigates.

Install the Gremlin Agent on the VMs backing your jobs and you can establish failure behavior in advance rather than reconstructing it afterward. With Gremlin, you can:

Find out whether interrupted jobs resume or restart. Shutdown experiments remove a VM mid-execution so you can watch what Batch does next, and how much compute you lose.
Catch partial output before it reaches production. Process Killer experiments end a job process partway through, revealing whether your pipeline writes incomplete results that look finished.
Test scratch space limits. Disk experiments fill local storage during a run, which is where jobs fail in ways their authors never anticipated.

For pipelines feeding reporting or downstream processing, knowing whether interrupted output can be trusted is worth more than knowing the job usually succeeds.

Building resilience on Batch (VM-backed jobs) with Gremlin

Batch (VM-backed jobs) is an infrastructure service, which means it supports the Gremlin agent *. This lets you run the full suite of Gremlin experiments. Using the Gremlin agent, you can find deep, infrastructure-level reliability risks in your Batch (VM-backed jobs) workloads, such as:

Scaling problems due to missing or misconfigured autoscaling rules
Processes that don't auto-recover after a crash
Limited redundancy on the host and network level

You can also run these expert-built workflows designed to replicate real-world failure modes on Batch (VM-backed jobs):

Host failure - Linux

Prepare for Linux host failures by shutting down a percentage of your hosts and observing the results. Validate that your infrastructure handles host loss without impacting service availability.

Try it

Scalability: Disk I/O

Test that your service scales as expected when Disk I/O is limited. Gremlin will carry out many read and write operations in the target's /var/tmp directory. Estimated test length: 20 minutes.

Try it

Host redundancy - Linux

Test resilience to host failures by shutting down a randomly selected Linux host. Verify that your platform automatically restarts or replaces it.

Try it

* Gremlin is designed to work on any cloud platform that provides Linux or Windows hosts. We haven't individually tested every service we cover, and not all are officially supported. Check our compatibility documentation for tested operating systems and known caveats, or get in touch if you don't see yours.

Learn more about Gremlin and Batch (VM-backed jobs)

Intelligent Health Checks - GCP

Automatically monitor your Google Cloud services during testing in one-click with Intelligent Health Checks.

Read

All product names, logos, and brands are property of their respective owners. AWS is a trademark of Amazon.com, Inc.; Azure is a trademark of Microsoft Corporation; Google Cloud is a trademark of Google LLC. Use of these names is for identification purposes only and does not imply endorsement or affiliation unless otherwise stated.

Avoid downtime. Use Gremlin to turn failure into resilience.

Gremlin empowers you to proactively root out failure before it causes downtime. See how you can harness chaos to build resilient systems by requesting a demo of Gremlin.

Get a demo

Gremlin and GCP Batch (VM-backed jobs)

Test GCP Batch reliability with Gremlin. Simulate process crashes and VM failures to validate job rescheduling.

Why Batch (VM-backed jobs) reliability is important

Install the Gremlin Agent on the VMs backing your jobs and you can establish failure behavior in advance rather than reconstructing it afterward. With Gremlin, you can:

Find out whether interrupted jobs resume or restart. Shutdown experiments remove a VM mid-execution so you can watch what Batch does next, and how much compute you lose.
Catch partial output before it reaches production. Process Killer experiments end a job process partway through, revealing whether your pipeline writes incomplete results that look finished.
Test scratch space limits. Disk experiments fill local storage during a run, which is where jobs fail in ways their authors never anticipated.

For pipelines feeding reporting or downstream processing, knowing whether interrupted output can be trusted is worth more than knowing the job usually succeeds.

Building resilience on Batch (VM-backed jobs) with Gremlin

Scaling problems due to missing or misconfigured autoscaling rules
Processes that don't auto-recover after a crash
Limited redundancy on the host and network level

You can also run these expert-built workflows designed to replicate real-world failure modes on Batch (VM-backed jobs):

resources

Learn more about Gremlin and Batch (VM-backed jobs)

Avoid downtime. Use Gremlin to turn failure into resilience.

Gremlin empowers you to proactively root out failure before it causes downtime. See how you can harness chaos to build resilient systems by requesting a demo of Gremlin.

Get a demo