feat(deployments): scaffold plugin API and registry (AIRCORE-755) - #280
Merged
tylersbray merged 8 commits intoJun 22, 2026
Merged
Conversation
![github-advanced-security[bot]](https://avatars.githubusercontent.com/in/57789?s=60&v=4){kind=small}
Contributor
|
Contributor
|
Note Reviews pausedIt looks like this branch is under active development. To avoid overwhelming you with review comments due to an influx of new commits, CodeRabbit has automatically paused this review. You can configure this behavior by changing the Use the following commands to manage reviews:
Use the checkboxes below for quick actions:
📝 WalkthroughWalkthroughAdds the ChangesNeMo Deployments Plugin
Suggested reviewers
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Actionable comments posted: 10
🧹 Nitpick comments (3)
plugins/nemo-deployments/README.md (1)
3-4: 💤 Low valueClarify "deployment lifecycle" with concrete scope.
Replace the vague term "deployment lifecycle" with specific capabilities: e.g., "Create, read, update, and delete deployment configurations and active deployments; manage volumes; track deployment status."
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@plugins/nemo-deployments/README.md` around lines 3 - 4, Update the README description to replace the vague phrase "deployment lifecycle" with a concrete list of supported capabilities: state that the plugin provides create/read/update/delete (CRUD) for deployment configurations and active deployments, volume management, deployment status tracking, entity schemas, CRUD APIs, a DeploymentBackend abstract base class, and an executor registry; ensure the one-line summary in the README (top paragraph) explicitly enumerates these capabilities so readers immediately see the plugin's scope.Source: Coding guidelines
plugins/nemo-deployments/src/nemo_deployments_plugin/entities.py (1)
243-247: ⚡ Quick winNo-op validator is misleading.
This
@model_validatordoes not validate anything and always returnsself, so it implies an invariant that is not enforced. Convert this to a plain code comment/doc note, or implement an actual check where bothDeploymentand itsDeploymentConfig.restart_policyare available.Suggested cleanup
- `@model_validator`(mode="after") - def _document_status_restart_policy_consistency(self) -> Deployment: - """Document restart_policy vs terminal status expectations for the reconciler.""" - # Reconciler enforces: Never → SUCCEEDED terminal; Always/OnFailure → READY while running. - return self + # Restart-policy/terminal-status consistency is enforced by the reconciler + # when DeploymentConfig is resolved for this Deployment.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@plugins/nemo-deployments/src/nemo_deployments_plugin/entities.py` around lines 243 - 247, The _document_status_restart_policy_consistency method is a no-op model_validator that misleads readers; either remove the `@model_validator` and convert the docstring into a plain class/module comment, or implement an actual validation that accesses self.status and self.deployment_config.restart_policy (or DeploymentConfig.restart_policy) and raises a ValueError/ValidationError when the pair is inconsistent (e.g., restart_policy == "Never" requires terminal SUCCEEDED state, restart_policy in {"Always","OnFailure"} requires non-terminal/READY while running). Update or remove the decorator accordingly and keep the descriptive text explaining the reconciler expectations.plugins/nemo-deployments/tests/unit/test_registry.py (1)
35-53: ⚡ Quick winUse concrete return types in
_StubBackendinstead ofAny.These methods can return concrete
LogResult/VolumeStatusUpdatetypes directly; keepingAnyhere weakens contract checking in tests.As per coding guidelines, "Always prefer concrete type hints over string-based ones; do not import types under TYPE_CHECKING, instead prefer regular imports when possible."
Proposed diff
from typing import Any import pytest -from nemo_deployments_plugin.backends.abc import BackendStatusUpdate, DeploymentBackend +from nemo_deployments_plugin.backends.abc import ( + BackendStatusUpdate, + DeploymentBackend, + LogResult, + VolumeStatusUpdate, +) @@ - async def get_logs(self, **kwargs: Any) -> Any: - from nemo_deployments_plugin.backends.abc import LogResult - + async def get_logs(self, **kwargs: Any) -> LogResult: return LogResult(lines=[]) @@ - async def create_volume(self, **kwargs: Any) -> Any: - from nemo_deployments_plugin.backends.abc import VolumeStatusUpdate - + async def create_volume(self, **kwargs: Any) -> VolumeStatusUpdate: return VolumeStatusUpdate(status="PENDING") @@ - async def read_volume_status(self, **kwargs: Any) -> Any: - from nemo_deployments_plugin.backends.abc import VolumeStatusUpdate - + async def read_volume_status(self, **kwargs: Any) -> VolumeStatusUpdate: return VolumeStatusUpdate(status="BOUND") @@ - async def delete_volume(self, workspace: str, name: str) -> Any: - from nemo_deployments_plugin.backends.abc import VolumeStatusUpdate - + async def delete_volume(self, workspace: str, name: str) -> VolumeStatusUpdate: return VolumeStatusUpdate(status="RELEASED")🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@plugins/nemo-deployments/tests/unit/test_registry.py` around lines 35 - 53, Update _StubBackend methods to use concrete return types instead of Any: change signatures of get_logs to return nemo_deployments_plugin.backends.abc.LogResult and create_volume/read_volume_status/delete_volume to return nemo_deployments_plugin.backends.abc.VolumeStatusUpdate; add regular imports for LogResult and VolumeStatusUpdate at top of the test file (do not use TYPE_CHECKING or string types) and return the concrete instances as already constructed in the method bodies so the type hints match the returned objects.Source: Coding guidelines
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@plugins/nemo-deployments/pyproject.toml`:
- Around line 7-11: The dependencies list in pyproject.toml for the
nemo-deployments plugin is missing fastapi, causing runtime ModuleNotFoundError
when plugin API modules import it; update the [dependencies] block (the
dependencies = [...] list in pyproject.toml for the nemo-deployments package) to
include "fastapi" (optionally with a suitable version constraint, e.g. >=0.95)
so fastapi is installed at runtime alongside "nemo-platform",
"nemo-platform-plugin", and "pydantic>=2.10.6".
In `@plugins/nemo-deployments/README.md`:
- Around line 1-5: Add a top-level "Prerequisites" section before the existing
"# NeMo Deployments Plugin" content that lists required software and setup steps
(e.g., supported Python version, required package manager such as uv, NeMo
Platform runtime availability/configuration, any required environment variables
or credentials, and optional tooling like Docker/uv CLI if applicable); update
README.md to include this section title "Prerequisites" and concise bullet
points so readers know what must be installed or configured before continuing.
- Around line 15-18: Add a "Next Steps" section at the end of
plugins/nemo-deployments/README.md titled "Next Steps" that provides cross-links
to the plugin's API reference, the backend implementation guide (reference
AIRCORE-756, AIRCORE-757, AIRCORE-758), and any other related plugin
documentation; place it immediately after the setup/test commands shown (the uv
sync / pytest block), include one-line descriptions for each link (what the
reader will find there), and ensure links are relative or to the canonical docs
site consistent with other plugin READMEs.
- Around line 1-18: The README mixes reference, how-to, and explanatory content;
split it into a single-quadrant page or reorganize with progressive disclosure.
Choose either: (A) convert this README into a HOW-TO "Set up and test the
plugin" that keeps the uv sync/pytest commands and a brief 30s summary, then
move the API base path and architecture/Backend/DeploymentBackend ABC details
into separate REFERENCE and EXPLANATION pages; or (B) expand this README into
layered sections (Layer 1: 30s summary of purpose/value and intended audience;
Layer 2: 3–5min core concepts and mention of DeploymentBackend and executor
registry; Layer 3: 10min+ API endpoints with the
`/apis/deployments/v1/workspaces/{workspace}/...` example and cross-workspace
sentinel `-`; Layer 4: links to separate reference pages for full API spec and
schema). Update headings accordingly (e.g., "How to set up and test",
"Concepts", "API reference", "Further reading") and move details about the
DeploymentBackend ABC and executor registry into dedicated files or sections to
avoid mixing quadrants.
In `@plugins/nemo-deployments/src/nemo_deployments_plugin/api/v1/dependencies.py`:
- Around line 17-24: The code currently lets X-NMP-Principal-On-Behalf-Of
override X-NMP-Principal-Id via _effective_principal_id, allowing spoofing;
change require_service_principal to read the principal directly from
request.headers.get(_PRINCIPAL_ID_HEADER, "") (or alter _effective_principal_id
to not consult _ON_BEHALF_OF_HEADER) and perform the startswith("service:")
check against that trustworthy value (use the symbols _PRINCIPAL_ID_HEADER,
_ON_BEHALF_OF_HEADER, _effective_principal_id, and require_service_principal to
locate and update the logic).
In `@plugins/nemo-deployments/src/nemo_deployments_plugin/api/v1/deployments.py`:
- Around line 140-143: The update call to entity_client.update(deployment) can
raise NemoEntityConflictError during concurrent writes; modify the try/except
around that call to also catch NemoEntityConflictError (in addition to
NemoEntityNotFoundError) and return an HTTP 409 conflict response (matching the
status-update endpoints) with a concise message including the deployment name,
while keeping the existing logger.info for NemoEntityNotFoundError; locate the
entity_client.update(deployment) call and add the new except
NemoEntityConflictError branch to produce the 409 response.
In `@plugins/nemo-deployments/src/nemo_deployments_plugin/backends/registry.py`:
- Around line 56-62: The registry construction loop in the classmethod that
builds executors is not failure-atomic: if one backend instantiation
(classes[spec.backend](sdk, spec.config)) raises, previously-created backend
instances remain running. Wrap the executor creation loop in a try/except, and
on any exception iterate over the already-created executors (executors.values())
and call their shutdown/close method (e.g., shutdown() or close(), whichever the
backend implements) inside its own try/except to swallow secondary errors, then
re-raise the original exception; keep the existing UnknownBackendTypeError check
and final return of cls(executors, default_executor=default_executor).
In `@plugins/nemo-deployments/src/nemo_deployments_plugin/config.py`:
- Around line 32-33: Add a validation step to the config model that defines
port_range_start and port_range_end to enforce 1 <= port_range_start <= 65535, 1
<= port_range_end <= 65535 and port_range_start <= port_range_end at load time;
implement this as a Pydantic validator (e.g., a `@root_validator` or two
`@validator` methods) on the config class that contains the port_range_start and
port_range_end fields so it raises a clear ValueError when bounds or ordering
are invalid, preventing downstream runtime failures.
In `@plugins/nemo-deployments/src/nemo_deployments_plugin/schema.py`:
- Line 77: The restart_policy field in the schema currently uses a loose type
(restart_policy: str | None = None); change it to use the RestartPolicy enum
type (e.g., restart_policy: RestartPolicy | None = None) so only valid policies
are accepted; import or reference the RestartPolicy enum (or class) used
elsewhere in the repo and update any related validation or type annotations in
the same module (schema.py) to use RestartPolicy instead of raw str.
In `@plugins/nemo-deployments/tests/unit/test_api_volumes.py`:
- Around line 32-39: The test should assert the object passed into
mock_entity_client.create to ensure the route sets Volume.status="PENDING"
before persisting; in test_create_volume_201 capture the awaited call to
mock_entity_client.create (e.g., inspect mock_entity_client.create.await_args or
use assert_awaited_once and read call args) and assert the passed volume object
has status == "PENDING" (and optionally name/size match the request) rather than
only asserting the mocked return value.
---
Nitpick comments:
In `@plugins/nemo-deployments/README.md`:
- Around line 3-4: Update the README description to replace the vague phrase
"deployment lifecycle" with a concrete list of supported capabilities: state
that the plugin provides create/read/update/delete (CRUD) for deployment
configurations and active deployments, volume management, deployment status
tracking, entity schemas, CRUD APIs, a DeploymentBackend abstract base class,
and an executor registry; ensure the one-line summary in the README (top
paragraph) explicitly enumerates these capabilities so readers immediately see
the plugin's scope.
In `@plugins/nemo-deployments/src/nemo_deployments_plugin/entities.py`:
- Around line 243-247: The _document_status_restart_policy_consistency method is
a no-op model_validator that misleads readers; either remove the
`@model_validator` and convert the docstring into a plain class/module comment, or
implement an actual validation that accesses self.status and
self.deployment_config.restart_policy (or DeploymentConfig.restart_policy) and
raises a ValueError/ValidationError when the pair is inconsistent (e.g.,
restart_policy == "Never" requires terminal SUCCEEDED state, restart_policy in
{"Always","OnFailure"} requires non-terminal/READY while running). Update or
remove the decorator accordingly and keep the descriptive text explaining the
reconciler expectations.
In `@plugins/nemo-deployments/tests/unit/test_registry.py`:
- Around line 35-53: Update _StubBackend methods to use concrete return types
instead of Any: change signatures of get_logs to return
nemo_deployments_plugin.backends.abc.LogResult and
create_volume/read_volume_status/delete_volume to return
nemo_deployments_plugin.backends.abc.VolumeStatusUpdate; add regular imports for
LogResult and VolumeStatusUpdate at top of the test file (do not use
TYPE_CHECKING or string types) and return the concrete instances as already
constructed in the method bodies so the type hints match the returned objects.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Enterprise
Run ID: 13375584-28cd-4917-9bf5-7df481d974cb
⛔ Files ignored due to path filters (1)
uv.lockis excluded by!**/*.lock
📒 Files selected for processing (25)
plugins/nemo-deployments/README.mdplugins/nemo-deployments/pyproject.tomlplugins/nemo-deployments/src/nemo_deployments_plugin/api/v1/dependencies.pyplugins/nemo-deployments/src/nemo_deployments_plugin/api/v1/deployment_configs.pyplugins/nemo-deployments/src/nemo_deployments_plugin/api/v1/deployments.pyplugins/nemo-deployments/src/nemo_deployments_plugin/api/v1/status.pyplugins/nemo-deployments/src/nemo_deployments_plugin/api/v1/volumes.pyplugins/nemo-deployments/src/nemo_deployments_plugin/backends/abc.pyplugins/nemo-deployments/src/nemo_deployments_plugin/backends/registry.pyplugins/nemo-deployments/src/nemo_deployments_plugin/config.pyplugins/nemo-deployments/src/nemo_deployments_plugin/constants.pyplugins/nemo-deployments/src/nemo_deployments_plugin/entities.pyplugins/nemo-deployments/src/nemo_deployments_plugin/schema.pyplugins/nemo-deployments/src/nemo_deployments_plugin/service.pyplugins/nemo-deployments/src/nemo_deployments_plugin/validation.pyplugins/nemo-deployments/tests/unit/helpers.pyplugins/nemo-deployments/tests/unit/test_api_deployment_configs.pyplugins/nemo-deployments/tests/unit/test_api_deployments.pyplugins/nemo-deployments/tests/unit/test_api_status.pyplugins/nemo-deployments/tests/unit/test_api_volumes.pyplugins/nemo-deployments/tests/unit/test_entities.pyplugins/nemo-deployments/tests/unit/test_prerequisite_validation.pyplugins/nemo-deployments/tests/unit/test_registry.pyplugins/nemo-deployments/tests/unit/test_service_startup.pypyproject.toml
tylersbray
changed the title
4 tasks
tylersbray
added a commit
that referenced
this pull request
Jun 12, 2026
Introduce DeploymentsController with deployment/volume reconcilers, prerequisite gating, drift recovery, and orphan cleanup on top of the 755 plugin scaffold. Stacks on PR #280 (AIRCORE-755). AIRCORE-758 Signed-off-by: Tyler Bray <tbray@nvidia.com>
benmccown
requested changes
Jun 12, 2026
benmccown
left a comment
benmccown
left a comment
Contributor
There was a problem hiding this comment.
Mostly minor stuff that an agent should be able to address without too much issue. Nice work so far.
tylersbray
added a commit
that referenced
this pull request
Jun 15, 2026
Move API to v2, trim README, remove status_history from status PUT, drop plugin-level port range, and add referential delete guards for configs and volumes. Signed-off-by: Tyler Bray <tbray@nvidia.com>
Contributor
There was a problem hiding this comment.
Actionable comments posted: 1
♻️ Duplicate comments (1)
plugins/nemo-deployments/tests/unit/test_api_volumes.py (1)
33-44:⚠️ Potential issue | 🟡 Minor | ⚡ Quick winAssert persisted
created.statusin create-volume test.This is still missing from the prior review point: the response can stay
"PENDING"via mock return even if route stops setting status before persistence. Assertmock_entity_client.createpayload status directly.Patch
def test_create_volume_201(client: TestClient, mock_entity_client: AsyncMock) -> None: mock_entity_client.create.return_value = make_volume() resp = client.post( "/apis/deployments/v2/workspaces/default/volumes", json={"name": "vol1", "size": "5Gi"}, ) assert resp.status_code == 201 assert resp.json()["status"] == "PENDING" + mock_entity_client.create.assert_awaited_once() created = mock_entity_client.create.await_args.args[0] + assert created.status == "PENDING" assert created.name == "vol1" assert created.size == "5Gi" assert created.workspace == "default"🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@plugins/nemo-deployments/tests/unit/test_api_volumes.py` around lines 33 - 44, The test_create_volume_201 function is missing an assertion to verify the status field on the created volume object that gets persisted. Currently the test asserts the response status is "PENDING" but does not verify that the actual payload being passed to mock_entity_client.create() has the correct status value set. Add an assertion after the existing assertions to verify that created.status equals the expected status value (likely "PENDING"). This ensures the route is properly setting the status on the volume object before persisting it, not relying on the mock's return value to hide a missing status assignment.
🧹 Nitpick comments (1)
plugins/nemo-deployments/README.md (1)
1-11: ⚡ Quick winREADME omits required prerequisites and next steps sections per coding guidelines.
Current file violates two mandatory guidelines:
- "Always list prerequisites at the top of documentation pages before other content" — missing prerequisites section.
- "Include 'Next Steps' section at the end with cross-links to related documentation content" — missing next steps section.
This reflects the maintainer's request to simplify during scaffolding, but guidelines require both sections. Resolve this tension with maintainer (benmccown) — either add minimal prerequisites (Python version, uv, NeMo Platform availability) and next steps (links to backend implementation PRs AIRCORE-756/757/758), or document in issue that these are deferred until backends are ready.
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@plugins/nemo-deployments/README.md` around lines 1 - 11, The README.md file for the NeMo Deployments Plugin is missing two mandatory sections required by coding guidelines: a Prerequisites section at the top (before the current content) and a Next Steps section at the end. Add a Prerequisites section immediately after the main heading that lists Python version requirements, the uv tool dependency, and NeMo Platform availability. Add a Next Steps section at the end of the file with cross-links to related documentation and the backend implementation PRs (AIRCORE-756, AIRCORE-757, AIRCORE-758) to guide users on what to explore next. If there is uncertainty about the specific content, coordinate with the maintainer (benmccown) to either confirm the exact prerequisites and next steps to include, or document in an issue that these sections are deferred pending backend readiness.Source: Coding guidelines
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In
`@plugins/nemo-deployments/src/nemo_deployments_plugin/api/v2/deployment_configs.py`:
- Around line 132-147: The deployment config deletion is not atomic because the
check for referencing deployments (via deployment_names_using_config) and the
actual deletion (via entity_client.delete) happen in separate calls, creating a
race condition where a concurrent deployment creation could reference a config
after it passes the check but before deletion completes. Refactor the code to
use a single storage-level conditional delete operation that atomically verifies
no referencing deployments exist and deletes the DeploymentConfig in one call,
or implement an equivalent compare-and-swap contract to ensure the "no
referencing deployments" state and deletion occur together without allowing
concurrent operations to insert references in between.
---
Duplicate comments:
In `@plugins/nemo-deployments/tests/unit/test_api_volumes.py`:
- Around line 33-44: The test_create_volume_201 function is missing an assertion
to verify the status field on the created volume object that gets persisted.
Currently the test asserts the response status is "PENDING" but does not verify
that the actual payload being passed to mock_entity_client.create() has the
correct status value set. Add an assertion after the existing assertions to
verify that created.status equals the expected status value (likely "PENDING").
This ensures the route is properly setting the status on the volume object
before persisting it, not relying on the mock's return value to hide a missing
status assignment.
---
Nitpick comments:
In `@plugins/nemo-deployments/README.md`:
- Around line 1-11: The README.md file for the NeMo Deployments Plugin is
missing two mandatory sections required by coding guidelines: a Prerequisites
section at the top (before the current content) and a Next Steps section at the
end. Add a Prerequisites section immediately after the main heading that lists
Python version requirements, the uv tool dependency, and NeMo Platform
availability. Add a Next Steps section at the end of the file with cross-links
to related documentation and the backend implementation PRs (AIRCORE-756,
AIRCORE-757, AIRCORE-758) to guide users on what to explore next. If there is
uncertainty about the specific content, coordinate with the maintainer
(benmccown) to either confirm the exact prerequisites and next steps to include,
or document in an issue that these sections are deferred pending backend
readiness.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Enterprise
Run ID: edb6aae1-ca29-4691-8e33-3c31648352d3
📒 Files selected for processing (16)
plugins/nemo-deployments/README.mdplugins/nemo-deployments/src/nemo_deployments_plugin/api/v2/dependencies.pyplugins/nemo-deployments/src/nemo_deployments_plugin/api/v2/deployment_configs.pyplugins/nemo-deployments/src/nemo_deployments_plugin/api/v2/deployments.pyplugins/nemo-deployments/src/nemo_deployments_plugin/api/v2/status.pyplugins/nemo-deployments/src/nemo_deployments_plugin/api/v2/volumes.pyplugins/nemo-deployments/src/nemo_deployments_plugin/config.pyplugins/nemo-deployments/src/nemo_deployments_plugin/references.pyplugins/nemo-deployments/src/nemo_deployments_plugin/schema.pyplugins/nemo-deployments/src/nemo_deployments_plugin/service.pyplugins/nemo-deployments/tests/unit/test_api_deployment_configs.pyplugins/nemo-deployments/tests/unit/test_api_deployments.pyplugins/nemo-deployments/tests/unit/test_api_status.pyplugins/nemo-deployments/tests/unit/test_api_volumes.pyplugins/nemo-deployments/tests/unit/test_references.pyplugins/nemo-deployments/tests/unit/test_service_startup.py
💤 Files with no reviewable changes (2)
- plugins/nemo-deployments/src/nemo_deployments_plugin/api/v2/dependencies.py
- plugins/nemo-deployments/src/nemo_deployments_plugin/schema.py
🚧 Files skipped from review as they are similar to previous changes (3)
- plugins/nemo-deployments/tests/unit/test_service_startup.py
- plugins/nemo-deployments/tests/unit/test_api_deployment_configs.py
- plugins/nemo-deployments/tests/unit/test_api_status.py
Contributor
|
Changes look good so far. Thoughts on the 3 remaining? Here's the final complete picture: Fully addressed (5/9 actionable items):
Not addressed (3/9 actionable items):
|
tylersbray
added a commit
that referenced
this pull request
Jun 22, 2026
Move API to v2, trim README, remove status_history from status PUT, drop plugin-level port range, and add referential delete guards for configs and volumes. Signed-off-by: Tyler Bray <tbray@nvidia.com>
tylersbray
added a commit
that referenced
this pull request
Jun 22, 2026
Rename CreateDeploymentRequest and Deployment entity field to deployment_config (with workspace/name ref parsing on create), and drop endpoints from the controller status PUT body. Signed-off-by: Tyler Bray <tbray@nvidia.com>
tylersbray
force-pushed
the
755-deployments-plugin-scaffold-the-plugin-entities-api-backend/tbray
branch
from
June 22, 2026 00:14
16b2cb9 to
4f7b14a
Compare
Contributor
Author
|
Thanks for the follow-up — quick status on the three:
Also rebased the branch onto current |
tylersbray
added a commit
that referenced
this pull request
Jun 22, 2026
Introduce DeploymentsController with deployment/volume reconcilers, prerequisite gating, drift recovery, and orphan cleanup on top of the 755 plugin scaffold. Stacks on PR #280 (AIRCORE-755). AIRCORE-758 Signed-off-by: Tyler Bray <tbray@nvidia.com>
benmccown
self-requested a review
June 22, 2026 19:58
benmccown
approved these changes
Jun 22, 2026
github-merge-queue
Bot
removed this pull request from the merge queue due to failed status checks
Jun 22, 2026
Add nemo-deployments-plugin as the substrate-agnostic contract layer for AIRCORE-755: DeploymentConfig/Deployment/Volume entities, v1 CRUD routes, controller status endpoints, DeploymentBackend ABC, and named executor registry. Plugin starts with zero backends; reconciler and docker/k8s backends land in follow-on tickets. Signed-off-by: Tyler Bray <tbray@nvidia.com>
Rename backends/abc.py to base.py and extract shared types to break the stdlib abc shadowing cycle. Fix auth to use Principal-Id only, add registry rollback and port-range validation, and wire pytest pythonpath for helpers. Signed-off-by: Tyler Bray <tbray@nvidia.com>
Use import abc to avoid self-import false positive and fail backends via init() hook so DeploymentBackend.__init__ is exercised in tests. Signed-off-by: Tyler Bray <tbray@nvidia.com>
Move API to v2, trim README, remove status_history from status PUT, drop plugin-level port range, and add referential delete guards for configs and volumes. Signed-off-by: Tyler Bray <tbray@nvidia.com>
Register get_authz_contribution on DeploymentsService so discovered routes satisfy static authz validation. Document referential delete TOCTOU limitation pending entity-store conditional delete. Signed-off-by: Tyler Bray <tbray@nvidia.com>
Rename CreateDeploymentRequest and Deployment entity field to deployment_config (with workspace/name ref parsing on create), and drop endpoints from the controller status PUT body. Signed-off-by: Tyler Bray <tbray@nvidia.com>
Rebase onto current main to pick up evaluator MetricFilter fix (#392) that caused merge-queue lint-openapi/lint-web-sdk failures. Commit the generated deployments plugin OpenAPI spec alongside other opted-in plugins. Signed-off-by: Tyler Bray <tbray@nvidia.com>
tylersbray
force-pushed
the
755-deployments-plugin-scaffold-the-plugin-entities-api-backend/tbray
branch
from
June 22, 2026 22:07
4f7b14a to
da124d3
Compare
Linux CI reuses ProcessPoolExecutor workers across plugins. Discovering services for one plugin imports route modules for others, registering query-param filter schemas at import time; the next plugin extraction clears that registry without re-importing routes, leaving dangling Filter refs (MetricFilter on evaluator). Use a fresh max_workers=1 pool per plugin, matching the services sequential-isolation pattern. Signed-off-by: Tyler Bray <tbray@nvidia.com>
tylersbray
deleted the
755-deployments-plugin-scaffold-the-plugin-entities-api-backend/tbray
branch
June 22, 2026 22:54
tylersbray
added a commit
that referenced
this pull request
Jun 22, 2026
Introduce DeploymentsController with deployment/volume reconcilers, prerequisite gating, drift recovery, and orphan cleanup on top of the 755 plugin scaffold. Stacks on PR #280 (AIRCORE-755). AIRCORE-758 Signed-off-by: Tyler Bray <tbray@nvidia.com>
4 tasks
tylersbray
added a commit
that referenced
this pull request
Jun 24, 2026
Introduce DeploymentsController with deployment/volume reconcilers, prerequisite gating, drift recovery, and orphan cleanup on top of the 755 plugin scaffold. Stacks on PR #280 (AIRCORE-755). AIRCORE-758 Signed-off-by: Tyler Bray <tbray@nvidia.com>
This was referenced Jun 25, 2026
7 tasks
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Scaffolds
nemo-deployments-pluginas the substrate-agnostic contract layer for AIRCORE-755:DeploymentConfig,Deployment,Volumeplus RFC supporting types/apis/deployments/v1/workspaces/{workspace}/...status_inlist filter; controller-onlyPUT .../status(service-principal gated viaX-NMP-Principal-Id)DeploymentBackendABC +BackendStatusUpdate/VolumeStatusUpdateNo substrate backends or reconcile controller in this PR — those are AIRCORE-756/757/758.
Test plan
uv run pytest plugins/nemo-deployments/tests/unit -v(31 tests)uv run ruff check plugins/nemo-deploymentsuv run ty check plugins/nemo-deploymentsno-nmp-common-in-plugins)nemo services run→ CRUD against/apis/deployments/v1/...(follow-on if desired)Summary by CodeRabbit
Release Notes
New Features
Tests
Documentation
Chores