feat(infrastructure): add private Linux Isaac Sim VM deployment option by fbeltrao · Pull Request #348 · microsoft/physical-ai-toolchain

fbeltrao · 2026-03-23T13:17:58Z

Description

Add an optional private Linux Isaac Sim VM deployment path for development environments.

This PR adds a Bicep-based VM deployment for the NVIDIA Isaac Sim Developer Workstation marketplace image, a deploy wrapper script that resolves values from the existing Terraform platform deployment, post-provisioning setup for workstation dependencies, and optional NAT egress and Microsoft Defender for Endpoint VM extensions. It also extends the Terraform platform outputs and variables so the deployment script can reuse a dedicated VM subnet and the shared network security group from the platform environment.

This is partial progress toward #63. The current implementation is intentionally limited to Linux VMs on private networking and reuses existing subnet and NSG resources instead of introducing a standalone Terraform VM deployment path.

What Is Included

Bicep entrypoint and modules for a private Linux Isaac Sim VM, optional subnet NAT egress, and shared default parameter types.
Deployment script support for Terraform-derived defaults, isolated VM resource groups, config preview, and optional Defender for Endpoint enablement.
Post-provisioning scripts for developer tooling and unattended ThinLinc installation.
Terraform updates that expose and optionally create the VM subnet and network security group needed by the deployment flow.
README documentation for prerequisites, deployment commands, parameters, and current limitations.

Not Included In This PR

Terraform-managed Isaac Sim VM resources. This PR adds Terraform support for the required network outputs only.
A standalone validation script such as scripts/validate_isaac_vm.sh.
Verification for the remaining issue acceptance criteria, including Checkov scanning, deployment time measurement, and automated Isaac Sim sample-scene launch validation.
Public IP or alternate connectivity modes. The current flow supports private networking only.
Additional authentication modes such as SSH key-based login.

Those items should be completed in follow-up PRs so this change can stay scoped to the initial private Linux deployment path.

Type of Change

🐛 Bug fix (non-breaking change fixing an issue)
✨ New feature (non-breaking change adding functionality)
💥 Breaking change (fix or feature causing existing functionality to change)
📚 Documentation update
🏗️ Infrastructure change (Terraform/IaC)
♻️ Refactoring (no functional changes)

Component(s) Affected

infrastructure/terraform/prerequisites/ - Azure subscription setup
infrastructure/terraform/ - Terraform infrastructure
infrastructure/setup/ - OSMO control plane / Helm
workflows/ - Training and evaluation workflows
training/ - Training pipelines and scripts
docs/ - Documentation

Testing Performed

Terraform plan reviewed (no unexpected changes)
Terraform apply tested in dev environment
Training scripts tested locally with Isaac Sim
OSMO workflow submitted successfully
Smoke tests passed (smoke_test_azure.py)

Notes:

terraform fmt was run for the Terraform changes.
./deploy-isaac-sim-vm.sh --vm-name <vm-name> was run successfully for the optional deployment flow.

Documentation Impact

No documentation changes needed
Documentation updated in this PR
Documentation issue filed

Issue Link

Partially addresses #63.

Checklist

My code follows the project conventions
Commit messages follow conventional commit format
I have performed a self-review
Documentation impact assessed above
No new linting warnings introduced

codecov-commenter · 2026-03-23T13:19:14Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 43.58%. Comparing base (9c34f86) to head (e896a08).
✅ All tests successful. No failed tests found.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #348   +/-   ##
=======================================
  Coverage   43.58%   43.58%           
=======================================
  Files         242      242           
  Lines       14840    14840           
  Branches     1855     1855           
=======================================
  Hits         6468     6468           
  Misses       8082     8082           
  Partials      290      290

Flag	Coverage Δ	*Carryforward flag
pester	`79.87% <ø> (ø)`
pytest	`6.89% <ø> (ø)`	Carriedforward from 6804630
pytest-dataviewer	`61.98% <ø> (ø)`
vitest	`50.72% <ø> (ø)`

*This pull request uses carry forward flags. Click here to find out more.

🚀 New features to boost your workflow:

📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

katriendg

PR Review — Isaac Sim VM Deployment

Nice work getting the initial private Linux deployment path scoped and working. This is a solid foundation for #63. Below is a summary of the inline review comments.

Findings Summary

#	Severity	File	Summary
1	💡 Note	`bicep/main.bicep`	Bicep acceptable for optional deployment; Terraform variant could be a follow-up
2	⚠️ High	`bicep/modules/subnet-nat-egress.bicep`	Subnet PUT overwrites existing NSG/delegation/route properties — carry forward existing config
3	⚠️ Medium	`scripts/install-dev-deps.sh`	Hardcoded `azureuser` for Docker group + `git config --global` runs as root — pass admin username from Bicep
4	💡 Suggestion	`deploy-isaac-sim-vm.sh`	Custom tfvars parser is fragile — `terraform console` alternative
5	⚠️ Medium	`isaac-sim-vm/README.md`	Prerequisites should spell out exact Terraform pipeline steps and commands
6	⚠️ Medium	`bicep/modules/linux-isaac-vm.bicep`	Auto-shutdown timezone hardcoded to W. Europe — parameterize with UTC default
7	⚠️ Medium	`isaac-sim-vm/README.md`	No cleanup/destroy documentation — users need explicit teardown guidance for Bicep-deployed resources

Sub-Issue Recommendation

Issue #63 has no sub-issues. This PR addresses roughly half the acceptance criteria. Recommend creating a sub-issue to track the remaining items:

Passes Checkov security scan
Deployment completes in under 15 minutes (measure and document)
Isaac Sim launches successfully and loads a sample scene after provisioning
Deployment validation script (scripts/validate_isaac_vm.sh)
Decision on Terraform variant requirement (or amend #63 criteria to Bicep-only)
SSH key authentication support (currently password-only)
Parameterize auto-shutdown timezone
Add cleanup/destroy documentation section to README
Make ThinLinc installation optional

…sions - add main Bicep file for Isaac Linux VM deployment - create module for Linux Isaac VM with necessary resources - implement subnet NAT egress module for outbound internet access - define shared types and default configurations for deployment - include scripts for installing development dependencies and ThinLinc silently 🔧 - Generated by Copilot

…schedule - add shutdown schedule parameter to Bicep modules - update deployment scripts to validate admin user existence - improve README with detailed deployment steps and cleanup instructions 🛠️ - Generated by Copilot

- specify that deleting the resource group will remove all VMs - add cautionary note for users regarding the command ⚠️ - Generated by Copilot

katriendg

Thanks for the additional changes, this looks good to me. The fact we have Bicep now only is OK, we can reflect on adding Terraform and enhancing any areas in subsequent PR.

🤖 I have created a release *beep* *boop* --- ## [0.5.0](v0.4.0...v0.5.0) (2026-03-26) ### ✨ Features * add dataviewer web application for dataset analysis and annotation ([#375](#375)) ([c44d7bb](c44d7bb)) * add return type annotations to cli_args functions ([#476](#476)) ([35523ee](35523ee)) * add YAML config schema with pydantic validation for ROS 2 recording ([#376](#376)) ([1fa5243](1fa5243)) * **agents:** Copilot agents and skills for dataviewer and OSMO training workflows. ([#444](#444)) ([8b72daf](8b72daf)) * **build:** add automated ms.date freshness checking ([#448](#448)) ([f92ddbc](f92ddbc)) * **build:** add CLA section, Dependabot security prefix, and OWASP ZAP DAST scan ([#241](#241)) ([083a8af](083a8af)) * **build:** add coverage.py configuration to pyproject.toml ([#428](#428)) ([eac7426](eac7426)) * **build:** add Go CI pipeline with golangci-lint and go test ([#351](#351)) ([b27e4fb](b27e4fb)) * **build:** add OpenSSF Scorecard workflow and badge ([#431](#431)) ([98a62e7](98a62e7)) * **build:** add release artifact signing and SBOM attestation ([#480](#480)) ([b226e96](b226e96)) * **build:** add TFLint reusable GitHub Actions workflow ([#229](#229)) ([34d5575](34d5575)) * **build:** split Go CI into separate lint and test pipelines ([#354](#354)) ([2dec155](2dec155)) * **dataviewer:** add authentication middleware and CSRF protection for mutation endpoints ([#432](#432)) ([77c8a01](77c8a01)) * **docs:** create training documentation hub with guides and migration ([#380](#380)) ([0fdccc5](0fdccc5)) * **docs:** port Docusaurus documentation site with full build validation ([#182](#182)) ([29dd640](29dd640)) * fix and deploy dataviewer ([#498](#498)) ([c922d49](c922d49)) * **inference:** add AzureML and local LeRobot inference workflows ([#438](#438)) ([f7d786a](f7d786a)) * **inference:** add MLflow trajectory plots and multi-source support to OSMO inference workflow ([#421](#421)) ([8637458](8637458)) * **infra:** add blob storage lifecycle policies and folder structure ([#179](#179)) ([101a6e8](101a6e8)) * **infrastructure:** add optional observability and compute feature flags ([#437](#437)) ([9eba0da](9eba0da)) * **infrastructure:** add private Linux Isaac Sim VM deployment option ([#348](#348)) ([3748c2d](3748c2d)) * **infrastructure:** add terraform-docs auto-generation pipeline ([#358](#358)) ([6565caa](6565caa)) * **infrastructure:** harden Isaac Sim VM deployment with encryption and spot options ([#355](#355)) ([6ebc1f2](6ebc1f2)) * **repo:** migrate to domain-driven architecture ([#270](#270)) ([a339e70](a339e70)) * **scripts:** add --config-preview and deployment summary to submission scripts ([#499](#499)) ([4069806](4069806)) * **scripts:** add Copilot attribution footer validation to frontmatter linting ([#378](#378)) ([4d595f2](4d595f2)) * **src:** add dataviewer web application with storage adapter layer ([#404](#404)) ([8a9fb70](8a9fb70)) ### 🐛 Bug Fixes * **build:** add GHSA to cspell custom dictionary ([#315](#315)) ([67db81a](67db81a)) * **build:** correct codecov report_type input for terraform test uploads ([#324](#324)) ([d90d66d](d90d66d)) * **build:** expand CODEOWNERS coverage to critical paths ([#505](#505)) ([bafade1](bafade1)) * **build:** pin Docker base image and pip dependencies with Dependabot coverage ([#497](#497)) ([d3d7ea4](d3d7ea4)) * **build:** pin pydantic version and use uv in config schema validation workflow ([#493](#493)) ([28d823f](28d823f)) * **build:** pin uv installer to versioned URL ([#495](#495)) ([8d8541b](8d8541b)) * **build:** remediate GHSA vulnerabilities flagged by OSSF Scorecard ([#271](#271)) ([49b6e58](49b6e58)) * **build:** remove README frontmatter, add FrontmatterExcludePaths, enforce Pester 5 ([#443](#443)) ([641d0f3](641d0f3)) * **build:** resolve CI failures for release 0.5.0 PR ([#174](#174)) ([62c9900](62c9900)) * **build:** resolve codecov PR comment suppression ([#523](#523)) ([5603bd7](5603bd7)) * **build:** use npm ci for deterministic frontend dependency install ([#491](#491)) ([ee8b5d3](ee8b5d3)), closes [#490](#490) * **ci:** add `wait_for_ci` to Codecov configuration ([#183](#183)) ([370cf44](370cf44)) * **CI:** Issue 116 clean up dataviewer tests ([#184](#184)) ([f466c23](f466c23)) * **ci:** pin pydantic to ==2.12.5 across all references ([#230](#230)) ([9d841d5](9d841d5)) * **dataviewer:** add HTTP Range support for blob video streaming ([#165](#165)) ([8adde50](8adde50)) * **dataviewer:** remediate CodeQL alerts and align ruff config ([#419](#419)) ([eb6fac9](eb6fac9)) * **dataviewer:** remediate path traversal and input validation vulnerabilities ([#413](#413)) ([0a1d2ca](0a1d2ca)) * **docs:** remove trailingSlash: false for GitHub Pages compatibility ([#228](#228)) ([a78cb97](a78cb97)) * **gpu:** add GPU Operator validation dependencies to GRID driver installer ([#441](#441)) ([eec42da](eec42da)) * **infrastructure:** add zone-redundant config to VPN gateway public IP ([#352](#352)) ([2d734f4](2d734f4)) * **infrastructure:** improve stdout handling for helm commands in GPU… ([#311](#311)) ([153f467](153f467)) * **infrastructure:** resolve remaining TFLint violations in SIL module and example configs ([#298](#298)) ([c0ce3e5](c0ce3e5)) * **infrastructure:** resolve TFLint violations in root and automation modules ([#287](#287)) ([b6a4604](b6a4604)), closes [#203](#203) * **infrastructure:** update deprecated bgp vng variable name ([#307](#307)) ([f530734](f530734)) * **scripts:** pin uv version in OSMO workflow templates ([#500](#500)) ([7edf13a](7edf13a)) * **scripts:** replace lambda with def in lerobot_handler to satisfy R… ([#176](#176)) ([baf9e58](baf9e58)) * **scripts:** support OSMO control-plane deploys with in-cluster Redis ([#317](#317)) ([d4b70de](d4b70de)) * **scripts:** update compute target name derivation logic ([#319](#319)) ([bb20431](bb20431)) * **settings:** update devcontainer name to match project context ([#177](#177)) ([745321e](745321e)) * **terraform:** create PostgreSQL Key Vault secret via ARM control plane ([#304](#304)) ([5d73b81](5d73b81)) * **terraform:** gate observability with feature flags ([#303](#303)) ([ea5e056](ea5e056)) * **terraform:** switch VPN gateway defaults to AZ SKUs ([#309](#309)) ([74989c5](74989c5)) * **training:** correct learning rate mapping and pin LeRobot version ([#439](#439)) ([5cf9943](5cf9943)) * **workflows:** enable SARIF upload for dependency-pinning scans ([#502](#502)) ([124cad6](124cad6)), closes [#501](#501) * **workflows:** remove redundant top-level permissions from codeql-analysis ([#489](#489)) ([1490fda](1490fda)) * **workflows:** use bash shell for uv.lock regeneration and add SARIF to dictionary ([#225](#225)) ([e6fa6ea](e6fa6ea)) ### 📚 Documentation * add chunking and compression configuration guide for Jetson edge recording ([#408](#408)) ([787a322](787a322)) * add OpenSSF Best Practices badge to README ([#282](#282)) ([01ea384](01ea384)) * add threat model cross-reference to SECURITY.md ([#235](#235)) ([88a461e](88a461e)) * add vulnerability remediation timeline to SECURITY.md ([#233](#233)) ([5ead3ee](5ead3ee)) * **contributing:** remove version-specific planning language from ownership tip ([#407](#407)) ([3191f9b](3191f9b)) * **deploy:** replace deploy/ READMEs with pointer files ([#379](#379)) ([b3c3abb](b3c3abb)) * **docs:** add bug report response timeline for OSSF report_responses criterion ([#485](#485)) ([9b26212](9b26212)) * **docs:** add component update process for OpenSSF Silver badge ([#446](#446)) ([6adc8a2](6adc8a2)) * **docs:** Add data collection and training recipes ([#343](#343)) ([9c34f86](9c34f86)) * **docs:** add deprecation policy for external interfaces ([#445](#445)) ([229d5db](229d5db)) * **docs:** add structure for recipes in repo ([#322](#322)) ([098757b](098757b)) * **docs:** add YAML frontmatter to SUPPORT.md ([#478](#478)) ([d94c15d](d94c15d)), closes [#347](#347) * **docs:** clarify issue assignment requirement before starting work ([#299](#299)) ([1534462](1534462)) * **docs:** create inference and training docs hubs ([#402](#402)) ([7a20a2e](7a20a2e)) * **docs:** create reference hub and migrate script documentation ([#503](#503)) ([03a31c6](03a31c6)) * **docs:** create training and inference documentation hubs ([#403](#403)) ([7be003b](7be003b)) * **operations:** create operations hub and troubleshooting guide ([#525](#525)) ([31c7aaa](31c7aaa)) * **reference:** add copilot artifacts documentation hub ([#170](#170)) ([9a45ca4](9a45ca4)) * simplify root README and update prerequisites ([#440](#440)) ([c0c7710](c0c7710)) ### ♻️ Code Refactoring * **build:** align Python dependency workflows with uv ([#447](#447)) ([3102e03](3102e03)) * **docs:** rename Docusaurus site to Physical AI Toolchain ([#224](#224)) ([cfdf47a](cfdf47a)) * **infrastructure:** rename boolean variables to `should_` prefix and add missing core variables ([#292](#292)) ([4496593](4496593)) * **python:** move runtime deps to workflow pyproject manifests ([#405](#405)) ([6c5fbeb](6c5fbeb)) ### 📦 Build System * **build:** add Codecov upload to pytest workflow ([#434](#434)) ([0110c17](0110c17)) * **deps-dev:** bump the npm_and_yarn group across 2 directories with 1 update ([#325](#325)) ([59cf9e6](59cf9e6)) * **workflows:** enable coverage parameters and fix Pester test infrastructure ([#435](#435)) ([528bbde](528bbde)) ### 🔧 Miscellaneous * add gomod to cspell general-technical wordlist ([#362](#362)) ([1f93f47](1f93f47)) * **build:** add codecov.yml for unified coverage reporting ([#430](#430)) ([b0faf70](b0faf70)) * **build:** add Go toolchain devcontainer feature and Dependabot gomod ([#337](#337)) ([8a36620](8a36620)) * **deps:** bump cryptography from 45.0.7 to 46.0.5 in /src/training ([#506](#506)) ([a06434e](a06434e)) * **deps:** bump minimatch in /src/dataviewer/frontend ([#416](#416)) ([38a7607](38a7607)) * **deps:** bump pyasn1 from 0.6.2 to 0.6.3 in /training/rl ([#296](#296)) ([7b42cf5](7b42cf5)) * **deps:** bump rollup in /src/dataviewer/frontend ([#417](#417)) ([6302ce4](6302ce4)) * **deps:** bump the common-dependencies group in /src/common with 3 updates ([#507](#507)) ([db05074](db05074)) * **deps:** bump the github-actions group across 1 directory with 6 updates ([#284](#284)) ([c40eff6](c40eff6)) * **deps:** bump the github-actions group across 1 directory with 6 updates ([#433](#433)) ([2d9dd4f](2d9dd4f)) * **deps:** bump the github-actions group across 1 directory with 6 updates ([#510](#510)) ([c334a64](c334a64)) * **deps:** bump the github-actions group with 2 updates ([#163](#163)) ([f25713e](f25713e)) * **deps:** bump the inference-dependencies group in /evaluation with 3 updates ([#279](#279)) ([1d2d3dc](1d2d3dc)) * **deps:** bump the inference-dependencies group in /src/inference with 5 updates ([#508](#508)) ([2852ffb](2852ffb)) * **deps:** bump the lerobot-inference-dependencies group in /workflows/azureml with 4 updates ([#511](#511)) ([b7c5773](b7c5773)) * **deps:** bump the npm_and_yarn group across 2 directories with 1 update ([#223](#223)) ([6a261ab](6a261ab)) * **deps:** bump the training-dependencies group ([#429](#429)) ([66e43f4](66e43f4)) * **deps:** bump tornado from 6.5.4 to 6.5.5 in the uv group across 1 directory ([#172](#172)) ([d6caf29](d6caf29)) * **docs:** correct ms.date tooling and refresh stale documentation ([#349](#349)) ([ccaa1e8](ccaa1e8)) * **infrastructure:** add Go module and golangci-lint config for e2e tests ([#347](#347)) ([e0e6bbf](e0e6bbf)) * **infrastructure:** add root .terraform-docs.yml configuration ([#312](#312)) ([bb73bbb](bb73bbb)) * migrate references from Azure-Samples to microsoft/physical-ai-toolchain ([f58f0ef](f58f0ef)) * **workflows:** update Dependabot, CodeQL, CODEOWNERS, and cspell for dataviewer coverage ([#231](#231)) ([6d8c2e8](6d8c2e8)) ### 🔒 Security * **deps:** bump mlflow from 3.5.0 to 3.8.0rc0 in /training/rl ([#297](#297)) ([e9929df](e9929df)) * **deps:** bump the github-actions group across 1 directory with 4 updates ([#344](#344)) ([6826929](6826929)) * **deps:** bump the inference-dependencies group in /evaluation with 2 updates ([#339](#339)) ([6804630](6804630)) * **deps:** bump the npm_and_yarn group across 3 directories with 1 update ([#361](#361)) ([6760857](6760857)) * **deps:** bump the training-dependencies group across 1 directory with 54 updates ([#286](#286)) ([d9ae04f](d9ae04f)) * **deps:** bump the uv group across 3 directories with 1 update ([#360](#360)) ([dfbda06](dfbda06)) --- This PR was generated with [Release Please](https://github.com/googleapis/release-please). See [documentation](https://github.com/googleapis/release-please#release-please). --------- Co-authored-by: physical-ai-toolchain-release[bot] <267194360+physical-ai-toolchain-release[bot]@users.noreply.github.com> Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Co-authored-by: Bill Berry <wbery@microsoft.com>

fbeltrao requested a review from a team as a code owner March 23, 2026 13:17

katriendg reviewed Mar 24, 2026

View reviewed changes

fbeltrao added 3 commits March 24, 2026 17:00

feat(cspell): add new Azure services to spell check list

a7580b1

fbeltrao force-pushed the feature/63-private-isaac-sim-linux branch from 66aa3da to dab4518 Compare March 24, 2026 17:00

fbeltrao added 2 commits March 24, 2026 17:16

docs(infrastructure): clarify resource group deletion warning in README

d99a95c

- specify that deleting the resource group will remove all VMs - add cautionary note for users regarding the command ⚠️ - Generated by Copilot

docs(infrastructure): update README for VM deployment steps formatting

4a0a642

katriendg approved these changes Mar 25, 2026

View reviewed changes

Merge branch 'main' into feature/63-private-isaac-sim-linux

e896a08

katriendg merged commit 3748c2d into microsoft:main Mar 25, 2026
27 checks passed

katriendg mentioned this pull request Mar 25, 2026

infra(evaluation): Isaac Sim Azure VM deployment template #63

Closed

8 tasks

physical-ai-toolchain-release Bot mentioned this pull request Mar 25, 2026

chore(main): release 0.5.0 #173

Merged

fbeltrao deleted the feature/63-private-isaac-sim-linux branch April 10, 2026 16:07

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

feat(infrastructure): add private Linux Isaac Sim VM deployment option#348

feat(infrastructure): add private Linux Isaac Sim VM deployment option#348
katriendg merged 6 commits into
microsoft:mainfrom
fbeltrao:feature/63-private-isaac-sim-linux

fbeltrao commented Mar 23, 2026

Uh oh!

codecov-commenter commented Mar 23, 2026 •

edited

Loading

Uh oh!

katriendg left a comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

katriendg left a comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Conversation

fbeltrao commented Mar 23, 2026

Description

What Is Included

Not Included In This PR

Type of Change

Component(s) Affected

Testing Performed

Documentation Impact

Issue Link

Checklist

Uh oh!

codecov-commenter commented Mar 23, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

katriendg left a comment

Choose a reason for hiding this comment

PR Review — Isaac Sim VM Deployment

Findings Summary

Sub-Issue Recommendation

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

katriendg left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

codecov-commenter commented Mar 23, 2026 •

edited

Loading