feat: use k8s fields for scheduling sessions

components/renku_data_services/crc/api.spec.yaml

@@ -550,7 +550,7 @@ paths:
           content:
             "application/json":
               schema:
-                $ref: "#/components/schemas/K8sLabelList"
+                $ref: "#/components/schemas/Tolerations"
         default:
           $ref: "#/components/responses/Error"
       tags:
@@ -1075,7 +1075,7 @@ components:
         default:
           $ref: "#/components/schemas/DefaultFlag"
         tolerations:
-          $ref: "#/components/schemas/K8sLabelList"
+          $ref: "#/components/schemas/Tolerations"
         node_affinities:
           $ref: "#/components/schemas/NodeAffinityList"
       required: ["cpu", "memory", "gpu", "max_storage", "name", "default", "default_storage"]
@@ -1106,7 +1106,7 @@ components:
         default:
           $ref: "#/components/schemas/DefaultFlagPatch"
         tolerations:
-          $ref: "#/components/schemas/K8sLabelList"
+          $ref: "#/components/schemas/Tolerations"
         node_affinities:
           $ref: "#/components/schemas/NodeAffinityList"
       example:
@@ -1133,7 +1133,7 @@ components:
         default:
           $ref: "#/components/schemas/DefaultFlagPatch"
         tolerations:
-          $ref: "#/components/schemas/K8sLabelList"
+          $ref: "#/components/schemas/Tolerations"
         node_affinities:
           $ref: "#/components/schemas/NodeAffinityList"
       required: ["id"]
@@ -1161,7 +1161,7 @@ components:
         default:
           $ref: "#/components/schemas/DefaultFlag"
         tolerations:
-          $ref: "#/components/schemas/K8sLabelList"
+          $ref: "#/components/schemas/Tolerations"
         node_affinities:
           $ref: "#/components/schemas/NodeAffinityList"
       required: ["cpu", "memory", "gpu", "max_storage", "name", "id", "default", "default_storage"]
@@ -1197,7 +1197,7 @@ components:
         matching:
           type: boolean
         tolerations:
-          $ref: "#/components/schemas/K8sLabelList"
+          $ref: "#/components/schemas/Tolerations"
         node_affinities:
           $ref: "#/components/schemas/NodeAffinityList"
       required: ["cpu", "memory", "gpu", "max_storage", "name", "id", "default", "default_storage"]
@@ -1800,15 +1800,65 @@ components:
       description: A threshold in seconds after which a session gets culled/deleted (0 means no threshold)
       minimum: 0
       maximum: 2147483647
-    K8sLabelList:
+    Tolerations:
       type: array
-      description: A list of k8s labels used for tolerations
+      description: |
+        A list of toleration items used for pod scheduling in Kubernetes.
       items:
-        $ref: "#/components/schemas/K8sLabel"
-      example: ["test-label-1"]
-      uniqueItems: true
+        $ref: "#/components/schemas/Toleration"
       default: []
-      minItems: 0
+    Toleration:
+      type: object
+      description: |
+        A toleration item.
+        The pod this Toleration is attached to tolerates any taint
+        that matches the triple <key,value,effect>
+        using the matching operator <operator>.
+      properties:
+        effect:
+          description: |-
+            Effect indicates the taint effect to match. Empty means match all taint effects.
+            When specified, allowed values are NoSchedule, PreferNoSchedule and NoExecute.
+          type: string
+        key:
+          description: |-
+            Key is the taint key that the toleration applies to. Empty means match all taint keys.
+            If the key is empty, operator must be Exists; this combination means to match all values and all keys.
+          type: string
+        operator:
+          description: |-
+            Operator represents a key's relationship to the value.
+            Valid operators are Exists and Equal. Defaults to Equal.
+            Exists is equivalent to wildcard for value, so that a pod can
+            tolerate all taints of a particular category.
+          type: string
+        tolerationSeconds:
+          description: |-
+            TolerationSeconds represents the period of time the toleration (which must be
+            of effect NoExecute, otherwise this field is ignored) tolerates the taint. By default,
+            it is not set, which means tolerate the taint forever (do not evict). Zero and
+            negative values will be treated as 0 (evict immediately) by the system.
+          format: int64
+          type: integer
+        value:
+          description: |-
+            Value is the taint value the toleration matches to.
+            If the operator is Exists, the value should be empty, otherwise just a regular string.
+          type: string
+      example:
+        effect: "NoSchedule"
+        key: "renku.io/dedicated"
+        operator: "Equal"
+        value: "user"
+    # K8sLabelList:
+    #   type: array
+    #   description: A list of k8s labels used for tolerations
+    #   items:
+    #     $ref: "#/components/schemas/K8sLabel"
+    #   example: ["test-label-1"]
+    #   uniqueItems: true
+    #   default: []
+    #   minItems: 0
     K8sLabel:
       type: string
       description: A valid K8s label

components/renku_data_services/crc/apispec.py

@@ -1,6 +1,6 @@
 # generated by datamodel-codegen:
 #   filename:  api.spec.yaml
-#   timestamp: 2025-10-13T09:06:36+00:00
+#   timestamp: 2025-10-31T14:40:52+00:00
 
 from __future__ import annotations
 
@@ -88,14 +88,26 @@ class RemoteConfigurationFirecrestPatch(BaseAPISpec):
     )
 
 
-class K8sLabel(RootModel[str]):
-    root: str = Field(
-        ...,
-        description="A valid K8s label",
-        examples=["some-label-1"],
-        max_length=63,
-        min_length=3,
-        pattern="^[a-z0-9A-Z][a-z0-9A-Z-_./]*[a-z0-9A-Z]$",
+class Toleration(BaseAPISpec):
+    effect: Optional[str] = Field(
+        None,
+        description="Effect indicates the taint effect to match. Empty means match all taint effects.\nWhen specified, allowed values are NoSchedule, PreferNoSchedule and NoExecute.",
+    )
+    key: Optional[str] = Field(
+        None,
+        description="Key is the taint key that the toleration applies to. Empty means match all taint keys.\nIf the key is empty, operator must be Exists; this combination means to match all values and all keys.",
+    )
+    operator: Optional[str] = Field(
+        None,
+        description="Operator represents a key's relationship to the value.\nValid operators are Exists and Equal. Defaults to Equal.\nExists is equivalent to wildcard for value, so that a pod can\ntolerate all taints of a particular category.",
+    )
+    tolerationSeconds: Optional[int] = Field(
+        None,
+        description="TolerationSeconds represents the period of time the toleration (which must be\nof effect NoExecute, otherwise this field is ignored) tolerates the taint. By default,\nit is not set, which means tolerate the taint forever (do not evict). Zero and\nnegative values will be treated as 0 (evict immediately) by the system.",
+    )
+    value: Optional[str] = Field(
+        None,
+        description="Value is the taint value the toleration matches to.\nIf the operator is Exists, the value should be empty, otherwise just a regular string.",
     )
 
 
@@ -485,11 +497,9 @@ class ResourceClass(BaseAPISpec):
         description="A default selection for resource classes or resource pools",
         examples=[False],
     )
-    tolerations: Optional[List[K8sLabel]] = Field(
+    tolerations: Optional[List[Toleration]] = Field(
         None,
-        description="A list of k8s labels used for tolerations",
-        examples=[["test-label-1"]],
-        min_length=0,
+        description="A list of toleration items used for pod scheduling in Kubernetes.\n",
     )
     node_affinities: Optional[List[NodeAffinity]] = Field(
         None,
@@ -541,11 +551,9 @@ class ResourceClassPatch(BaseAPISpec):
         description="A default selection for resource classes or resource pools",
         examples=[False],
     )
-    tolerations: Optional[List[K8sLabel]] = Field(
+    tolerations: Optional[List[Toleration]] = Field(
         None,
-        description="A list of k8s labels used for tolerations",
-        examples=[["test-label-1"]],
-        min_length=0,
+        description="A list of toleration items used for pod scheduling in Kubernetes.\n",
     )
     node_affinities: Optional[List[NodeAffinity]] = Field(
         None,
@@ -603,11 +611,9 @@ class ResourceClassPatchWithId(BaseAPISpec):
         description="A default selection for resource classes or resource pools",
         examples=[False],
     )
-    tolerations: Optional[List[K8sLabel]] = Field(
+    tolerations: Optional[List[Toleration]] = Field(
         None,
-        description="A list of k8s labels used for tolerations",
-        examples=[["test-label-1"]],
-        min_length=0,
+        description="A list of toleration items used for pod scheduling in Kubernetes.\n",
     )
     node_affinities: Optional[List[NodeAffinity]] = Field(
         None,
@@ -663,11 +669,9 @@ class ResourceClassWithId(BaseAPISpec):
         description="A default selection for resource classes or resource pools",
         examples=[False],
     )
-    tolerations: Optional[List[K8sLabel]] = Field(
+    tolerations: Optional[List[Toleration]] = Field(
         None,
-        description="A list of k8s labels used for tolerations",
-        examples=[["test-label-1"]],
-        min_length=0,
+        description="A list of toleration items used for pod scheduling in Kubernetes.\n",
     )
     node_affinities: Optional[List[NodeAffinity]] = Field(
         None,
@@ -724,11 +728,9 @@ class ResourceClassWithIdFiltered(BaseAPISpec):
         examples=[False],
     )
     matching: Optional[bool] = None
-    tolerations: Optional[List[K8sLabel]] = Field(
+    tolerations: Optional[List[Toleration]] = Field(
         None,
-        description="A list of k8s labels used for tolerations",
-        examples=[["test-label-1"]],
-        min_length=0,
+        description="A list of toleration items used for pod scheduling in Kubernetes.\n",
     )
     node_affinities: Optional[List[NodeAffinity]] = Field(
         None,

components/renku_data_services/crc/core.py

@@ -8,6 +8,9 @@
 from renku_data_services.base_models import RESET, ResetType
 from renku_data_services.crc import apispec, models
 from renku_data_services.errors import errors
+from renku_data_services.k8s.pod_scheduling import api as k8s_api
+from renku_data_services.k8s.pod_scheduling import models as k8s_models
+from renku_data_services.k8s.pod_scheduling import transforms as k8s_transforms
 
 
 def validate_quota(body: apispec.QuotaWithOptionalId) -> models.UnsavedQuota:
@@ -43,7 +46,11 @@ def validate_resource_class(body: apispec.ResourceClass) -> models.UnsavedResour
         ),
         key=lambda x: (x.key, x.required_during_scheduling),
     )
-    tolerations = sorted(t.root for t in body.tolerations or [])
+    api_tolerations = k8s_api.TolerationsField.model_validate(
+        [tol.model_dump(mode="json") for tol in body.tolerations or []]
+    )
+    tolerations = k8s_transforms.transform_tolerations(body=api_tolerations) or []
+    tolerations = sorted(tolerations, key=k8s_models.Toleration.sort_key)
     return models.UnsavedResourceClass(
         name=body.name,
         cpu=body.cpu,
@@ -91,9 +98,16 @@ def validate_resource_class_patch_or_put(
             ),
             key=lambda x: (x.key, x.required_during_scheduling),
         )
-    tolerations: list[str] | None = [] if method == "PUT" else None
+    # tolerations: list[str] | None = [] if method == "PUT" else None
+    # if body.tolerations:
+    #     tolerations = sorted(t.root for t in body.tolerations or [])
+    tolerations: list[k8s_models.Toleration] | None = [] if method == "PUT" else None
     if body.tolerations:
-        tolerations = sorted(t.root for t in body.tolerations or [])
+        api_tolerations = k8s_api.TolerationsField.model_validate(
+            [tol.model_dump(mode="json") for tol in body.tolerations]
+        )
+        tolerations = k8s_transforms.transform_tolerations(body=api_tolerations) or []
+        tolerations = sorted(tolerations, key=k8s_models.Toleration.sort_key)
     if rc_id:
         return models.ResourceClassPatchWithId(
             id=rc_id,

components/renku_data_services/crc/db.py

@@ -20,6 +20,7 @@
 
 import renku_data_services.base_models as base_models
 from renku_data_services import errors
+from renku_data_services.app_config import logging
 from renku_data_services.base_models import RESET
 from renku_data_services.crc import models
 from renku_data_services.crc import orm as schemas
@@ -29,6 +30,8 @@
 from renku_data_services.k8s.db import QuotaRepository
 from renku_data_services.users.db import UserRepo
 
+logger = logging.getLogger(__name__)
+
 
 class _Base:
     def __init__(self, session_maker: Callable[..., AsyncSession], quotas_repo: QuotaRepository) -> None:
@@ -338,8 +341,9 @@ async def get_classes(
                 # NOTE: The line below ensures that the right users can access the right resources, do not remove.
                 stmt = _classes_user_access_control(api_user, stmt)
 
-            res = await session.execute(stmt)
-            orms = res.scalars().all()
+            res = await session.scalars(stmt)
+            orms = res.all()
+            logger.warning(f"Classes = {[orm.dump() for orm in orms]}")
             return [orm.dump() for orm in orms]
 
     async def get_resource_class(self, api_user: base_models.APIUser, id: int) -> models.ResourceClass:
@@ -586,18 +590,31 @@ async def update_resource_class(
                         cls.node_affinities.remove(existing_affinity)
 
             if update.tolerations is not None:
-                existing_tolerations: dict[str, schemas.TolerationORM] = {tol.key: tol for tol in cls.tolerations}
-                new_tolerations: dict[str, schemas.TolerationORM] = {
-                    tol: schemas.TolerationORM(key=tol) for tol in update.tolerations
-                }
-                for new_tol_key, new_tol in new_tolerations.items():
-                    if new_tol_key not in existing_tolerations:
-                        # CREATE a brand new toleration
-                        cls.tolerations.append(new_tol)
-                # REMOVE a toleration
-                for existing_tol_key, existing_tol in existing_tolerations.items():
-                    if existing_tol_key not in new_tolerations:
-                        cls.tolerations.remove(existing_tol)
+                # existing_tolerations: dict[str, schemas.TolerationORM] = {tol.key: tol for tol in cls.tolerations}
+                # new_tolerations: dict[str, schemas.TolerationORM] = {
+                #     tol: schemas.TolerationORM(key=tol) for tol in update.tolerations
+                # }
+                # for new_tol_key, new_tol in new_tolerations.items():
+                #     if new_tol_key not in existing_tolerations:
+                #         # CREATE a brand new toleration
+                #         cls.tolerations.append(new_tol)
+                # # REMOVE a toleration
+                # for existing_tol_key, existing_tol in existing_tolerations.items():
+                #     if existing_tol_key not in new_tolerations:
+                #         cls.tolerations.remove(existing_tol)
+
+                # NOTE: the whole list of tolerations is updated
+                existing_tolerations = list(cls.new_tolerations)
+                for existing_tol, new_tol in zip(existing_tolerations, update.tolerations, strict=False):
+                    existing_tol.contents = new_tol.to_dict()
+
+                if len(update.tolerations) > len(existing_tolerations):
+                    # Add new tolerations
+                    for new_tol in update.tolerations[len(existing_tolerations) :]:
+                        cls.new_tolerations.append(schemas.NewTolerationORM.from_model(new_tol))
+                elif len(update.tolerations) < len(existing_tolerations):
+                    # Remove old tolerations
+                    cls.new_tolerations = cls.new_tolerations[: len(update.tolerations)]
 
             # NOTE: do we need to perform this check?
             if cls.resource_pool is None:

components/renku_data_services/crc/models.py

@@ -10,6 +10,7 @@
 from renku_data_services import errors
 from renku_data_services.base_models import ResetType
 from renku_data_services.k8s.constants import ClusterId
+from renku_data_services.k8s.pod_scheduling import models as k8s_models
 from renku_data_services.notebooks.cr_amalthea_session import TlsSecret
 
 
@@ -88,7 +89,8 @@ class UnsavedResourceClass(ResourcesCompareMixin):
     default: bool = False
     default_storage: int = 1
     node_affinities: list[NodeAffinity] = field(default_factory=list)
-    tolerations: list[str] = field(default_factory=list)
+    # tolerations: list[str] = field(default_factory=list)
+    tolerations: list[k8s_models.Toleration] = field(default_factory=list)
 
 
 @dataclass(frozen=True, eq=True, kw_only=True)
@@ -105,7 +107,8 @@ class ResourceClass(ResourcesCompareMixin):
     default_storage: int = 1
     matching: Optional[bool] = None
     node_affinities: list[NodeAffinity] = field(default_factory=list)
-    tolerations: list[str] = field(default_factory=list)
+    # tolerations: list[str] = field(default_factory=list)
+    tolerations: list[k8s_models.Toleration] = field(default_factory=list)
     quota: str | None = None
 
 
@@ -121,7 +124,8 @@ class ResourceClassPatch:
     default: bool | None = None
     default_storage: int | None = None
     node_affinities: list[NodeAffinity] | None = None
-    tolerations: list[str] | None = None
+    # tolerations: list[str] | None = None
+    tolerations: list[k8s_models.Toleration] | None = None
 
 
 @dataclass(frozen=True, eq=True, kw_only=True)