skypilot-nightly 1.0.0.dev20251027__py3-none-any.whl → 1.0.0.dev20251101__py3-none-any.whl

sky/provision/kubernetes/instance.py

@@ -33,6 +33,7 @@ from sky.utils.db import db_utils
 POLL_INTERVAL = 2
 _TIMEOUT_FOR_POD_TERMINATION = 60  # 1 minutes
 _MAX_RETRIES = 3
+_MAX_MISSING_PODS_RETRIES = 5
 _NUM_THREADS = subprocess_utils.get_parallel_threads('kubernetes')
 
 # Pattern to extract SSH user from command output, handling MOTD contamination
@@ -489,17 +490,17 @@ def _wait_for_pods_to_schedule(namespace, context, new_nodes, timeout: int,
 
 
 @timeline.event
-def _wait_for_pods_to_run(namespace, context, new_nodes):
+def _wait_for_pods_to_run(namespace, context, cluster_name, new_pods):
     """Wait for pods and their containers to be ready.
 
     Pods may be pulling images or may be in the process of container
     creation.
     """
-    if not new_nodes:
+    if not new_pods:
         return
 
     # Create a set of pod names we're waiting for
-    expected_pod_names = {node.metadata.name for node in new_nodes}
+    expected_pod_names = {pod.metadata.name for pod in new_pods}
 
     def _check_init_containers(pod):
         # Check if any of the init containers failed
@@ -526,28 +527,62 @@ def _wait_for_pods_to_run(namespace, context, new_nodes):
                     'Failed to create init container for pod '
                     f'{pod.metadata.name}. Error details: {msg}.')
 
+    missing_pods_retry = 0
     while True:
         # Get all pods in a single API call
-        cluster_name = new_nodes[0].metadata.labels[
+        cluster_name_on_cloud = new_pods[0].metadata.labels[
             k8s_constants.TAG_SKYPILOT_CLUSTER_NAME]
         all_pods = kubernetes.core_api(context).list_namespaced_pod(
             namespace,
             label_selector=
-            f'{k8s_constants.TAG_SKYPILOT_CLUSTER_NAME}={cluster_name}').items
+            f'{k8s_constants.TAG_SKYPILOT_CLUSTER_NAME}={cluster_name_on_cloud}'
+        ).items
 
         # Get the set of found pod names and check if we have all expected pods
         found_pod_names = {pod.metadata.name for pod in all_pods}
-        missing_pods = expected_pod_names - found_pod_names
-        if missing_pods:
+        missing_pod_names = expected_pod_names - found_pod_names
+        if missing_pod_names:
+            # In _wait_for_pods_to_schedule, we already wait for all pods to go
+            # from pending to scheduled. So if a pod is missing here, it means
+            # something unusual must have happened, and so should be treated as
+            # an exception.
+            # It is also only in _wait_for_pods_to_schedule that
+            # provision_timeout is used.
+            # TODO(kevin): Should we take provision_timeout into account here,
+            # instead of hardcoding the number of retries?
+            if missing_pods_retry >= _MAX_MISSING_PODS_RETRIES:
+                for pod_name in missing_pod_names:
+                    reason = _get_pod_missing_reason(context, namespace,
+                                                     cluster_name, pod_name)
+                    logger.warning(f'Pod {pod_name} missing: {reason}')
+                raise config_lib.KubernetesError(
+                    f'Failed to get all pods after {missing_pods_retry} '
+                    f'retries. Some pods may have been terminated or failed '
+                    f'unexpectedly. Run `sky logs --provision {cluster_name}` '
+                    'for more details.')
             logger.info('Retrying running pods check: '
-                        f'Missing pods: {missing_pods}')
+                        f'Missing pods: {missing_pod_names}')
             time.sleep(0.5)
+            missing_pods_retry += 1
             continue
 
         all_pods_running = True
         for pod in all_pods:
             if pod.metadata.name not in expected_pod_names:
                 continue
+
+            # Check if pod is terminated/preempted/failed.
+            if (pod.metadata.deletion_timestamp is not None or
+                    pod.status.phase == 'Failed'):
+                # Get the reason and write to cluster events before
+                # the pod gets completely deleted from the API.
+                reason = _get_pod_termination_reason(pod, cluster_name)
+                logger.warning(f'Pod {pod.metadata.name} terminated: {reason}')
+                raise config_lib.KubernetesError(
+                    f'Pod {pod.metadata.name} has terminated or failed '
+                    f'unexpectedly. Run `sky logs --provision {cluster_name}` '
+                    'for more details.')
+
             # Continue if pod and all the containers within the
             # pod are successfully created and running.
             if pod.status.phase == 'Running' and all(
@@ -1169,7 +1204,7 @@ def _create_pods(region: str, cluster_name: str, cluster_name_on_cloud: str,
     # fail early if there is an error
     logger.debug(f'run_instances: waiting for pods to be running (pulling '
                  f'images): {[pod.metadata.name for pod in pods]}')
-    _wait_for_pods_to_run(namespace, context, pods)
+    _wait_for_pods_to_run(namespace, context, cluster_name, pods)
     logger.debug(f'run_instances: all pods are scheduled and running: '
                  f'{[pod.metadata.name for pod in pods]}')
 
@@ -1428,9 +1463,45 @@ def get_cluster_info(
 
 
 def _get_pod_termination_reason(pod: Any, cluster_name: str) -> str:
-    """Get pod termination reason and write to cluster events."""
-    reasons = []
+    """Get pod termination reason and write to cluster events.
+
+    Checks both pod conditions (for preemption/disruption) and
+    container statuses (for exit codes/errors).
+    """
     latest_timestamp = pod.status.start_time or datetime.datetime.min
+    ready_state = 'Unknown'
+    termination_reason = 'Terminated unexpectedly'
+    container_reasons = []
+
+    # Check pod status conditions for high level overview.
+    # No need to sort, as each condition.type will only appear once.
+    for condition in pod.status.conditions:
+        reason = condition.reason or 'Unknown reason'
+        message = condition.message or ''
+
+        # Get last known readiness state.
+        if condition.type == 'Ready':
+            ready_state = f'{reason} ({message})' if message else reason
+        # Kueue preemption, as defined in:
+        # https://pkg.go.dev/sigs.k8s.io/kueue/pkg/controller/jobs/pod#pkg-constants
+        elif condition.type == 'TerminationTarget':
+            termination_reason = f'Preempted by Kueue: {reason}'
+            if message:
+                termination_reason += f' ({message})'
+        # Generic disruption.
+        elif condition.type == 'DisruptionTarget':
+            termination_reason = f'Disrupted: {reason}'
+            if message:
+                termination_reason += f' ({message})'
+
+        if condition.last_transition_time is not None:
+            latest_timestamp = max(latest_timestamp,
+                                   condition.last_transition_time)
+
+    pod_reason = (f'{termination_reason}.\n'
+                  f'Last known state: {ready_state}.')
+
+    # Check container statuses for exit codes/errors
     if pod.status and pod.status.container_statuses:
         for container_status in pod.status.container_statuses:
             terminated = container_status.state.terminated
@@ -1445,18 +1516,15 @@ def _get_pod_termination_reason(pod: Any, cluster_name: str) -> str:
                 if reason is None:
                     # just in-case reason is None, have default for debugging
                     reason = f'exit({exit_code})'
-                reasons.append(reason)
-                if terminated.finished_at > latest_timestamp:
-                    latest_timestamp = terminated.finished_at
+                container_reasons.append(reason)
+                latest_timestamp = max(latest_timestamp, terminated.finished_at)
 
             # TODO (kyuds): later, if needed, query `last_state` too.
 
-    if not reasons:
-        return ''
-
     # Normally we will have a single container per pod for skypilot
     # but doing this just in-case there are multiple containers.
-    pod_reason = ' | '.join(reasons)
+    if container_reasons:
+        pod_reason += f'\nContainer errors: {" | ".join(container_reasons)}'
 
     global_user_state.add_cluster_event(
         cluster_name,
@@ -1658,9 +1726,10 @@ def query_instances(
                                     Optional[str]]] = {}
     for pod in pods:
         phase = pod.status.phase
+        is_terminating = pod.metadata.deletion_timestamp is not None
         pod_status = status_map[phase]
         reason = None
-        if phase in ('Failed', 'Unknown'):
+        if phase in ('Failed', 'Unknown') or is_terminating:
             reason = _get_pod_termination_reason(pod, cluster_name)
             logger.debug(f'Pod Status ({phase}) Reason(s): {reason}')
         if non_terminated_only and pod_status is None:

sky/provision/kubernetes/volume.py

@@ -75,7 +75,6 @@ def delete_volume(config: models.VolumeConfig) -> models.VolumeConfig:
     """Deletes a volume."""
     context, namespace = _get_context_namespace(config)
     pvc_name = config.name_on_cloud
-    logger.info(f'Deleting PVC {pvc_name}')
     kubernetes_utils.delete_k8s_resource_with_retry(
         delete_func=lambda pvc_name=pvc_name: kubernetes.core_api(
             context).delete_namespaced_persistent_volume_claim(
@@ -84,6 +83,7 @@ def delete_volume(config: models.VolumeConfig) -> models.VolumeConfig:
                 _request_timeout=config_lib.DELETION_TIMEOUT),
         resource_type='pvc',
         resource_name=pvc_name)
+    logger.info(f'Deleted PVC {pvc_name} in namespace {namespace}')
     return config
 
 
@@ -242,9 +242,9 @@ def create_persistent_volume_claim(namespace: str, context: Optional[str],
     except kubernetes.api_exception() as e:
         if e.status != 404:  # Not found
             raise
-    logger.info(f'Creating PVC {pvc_name}')
     kubernetes.core_api(context).create_namespaced_persistent_volume_claim(
         namespace=namespace, body=pvc_spec)
+    logger.info(f'Created PVC {pvc_name} in namespace {namespace}')
 
 
 def _get_pvc_spec(namespace: str,

sky/schemas/api/responses.py

@@ -79,6 +79,7 @@ class APIHealthResponse(ResponseBaseModel):
     commit: str = ''
     basic_auth_enabled: bool = False
     user: Optional[models.User] = None
+    service_account_token_enabled: bool = False
 
 
 class StatusResponse(ResponseBaseModel):
@@ -90,7 +91,7 @@ class StatusResponse(ResponseBaseModel):
     # This is an internally facing field anyway, so it's less
     # of a problem that it's not typed.
     handle: Optional[Any] = None
-    last_use: str
+    last_use: Optional[str] = None
     status: status_lib.ClusterStatus
     autostop: int
     to_down: bool
@@ -98,11 +99,8 @@ class StatusResponse(ResponseBaseModel):
     # metadata is a JSON, so we use Any here.
     metadata: Optional[Dict[str, Any]] = None
     cluster_hash: str
-    # pydantic cannot generate the pydantic-core schema for
-    # storage_mounts_metadata, so we use Any here.
-    storage_mounts_metadata: Optional[Dict[str, Any]] = None
     cluster_ever_up: bool
-    status_updated_at: int
+    status_updated_at: Optional[int] = None
     user_hash: str
     user_name: str
     config_hash: Optional[str] = None

sky/schemas/db/spot_jobs/004_job_file_contents.py

@@ -0,0 +1,42 @@
+"""Add columns for stored DAG/env file contents.
+
+Revision ID: 004
+Revises: 003
+Create Date: 2025-10-27
+
+"""
+# pylint: disable=invalid-name
+from typing import Sequence, Union
+
+from alembic import op
+import sqlalchemy as sa
+
+from sky.utils.db import db_utils
+
+# revision identifiers, used by Alembic.
+revision: str = '004'
+down_revision: Union[str, Sequence[str], None] = '003'
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade():
+    """Add columns to persist job file contents in the database."""
+    with op.get_context().autocommit_block():
+        db_utils.add_column_to_table_alembic('job_info',
+                                             'dag_yaml_content',
+                                             sa.Text(),
+                                             server_default=None)
+        db_utils.add_column_to_table_alembic('job_info',
+                                             'original_user_yaml_content',
+                                             sa.Text(),
+                                             server_default=None)
+        db_utils.add_column_to_table_alembic('job_info',
+                                             'env_file_content',
+                                             sa.Text(),
+                                             server_default=None)
+
+
+def downgrade():
+    """No downgrade logic."""
+    pass

sky/schemas/db/spot_jobs/005_logs_gc.py

@@ -0,0 +1,38 @@
+"""Adding columns for the GC time of task logs and controller logs.
+
+Revision ID: 005
+Revises: 004
+Create Date: 2025-10-20
+
+"""
+# pylint: disable=invalid-name
+from typing import Sequence, Union
+
+from alembic import op
+import sqlalchemy as sa
+
+from sky.utils.db import db_utils
+
+# revision identifiers, used by Alembic.
+revision: str = '005'
+down_revision: Union[str, Sequence[str], None] = '004'
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade():
+    """Add columns for logs gc."""
+    with op.get_context().autocommit_block():
+        db_utils.add_column_to_table_alembic('job_info',
+                                             'controller_logs_cleaned_at',
+                                             sa.Float(),
+                                             server_default=None)
+        db_utils.add_column_to_table_alembic('spot',
+                                             'logs_cleaned_at',
+                                             sa.Float(),
+                                             server_default=None)
+
+
+def downgrade():
+    """Remove columns for logs gc."""
+    pass

sky/schemas/generated/managed_jobsv1_pb2.py

@@ -14,7 +14,7 @@ _sym_db = _symbol_database.Default()
 
 
 
-DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(b'\n*sky/schemas/generated/managed_jobsv1.proto\x12\x0fmanaged_jobs.v1\"\x15\n\x06JobIds\x12\x0b\n\x03ids\x18\x01 \x03(\x03\"\x1c\n\nUserHashes\x12\x0e\n\x06hashes\x18\x01 \x03(\t\"\x1c\n\x08Statuses\x12\x10\n\x08statuses\x18\x01 \x03(\t\"\x13\n\x11GetVersionRequest\"0\n\x12GetVersionResponse\x12\x1a\n\x12\x63ontroller_version\x18\x01 \x01(\t\"\xec\x03\n\x12GetJobTableRequest\x12\x15\n\rskip_finished\x18\x01 \x01(\x08\x12\x1d\n\x15\x61\x63\x63\x65ssible_workspaces\x18\x02 \x03(\t\x12-\n\x07job_ids\x18\x03 \x01(\x0b\x32\x17.managed_jobs.v1.JobIdsH\x00\x88\x01\x01\x12\x1c\n\x0fworkspace_match\x18\x04 \x01(\tH\x01\x88\x01\x01\x12\x17\n\nname_match\x18\x05 \x01(\tH\x02\x88\x01\x01\x12\x17\n\npool_match\x18\x06 \x01(\tH\x03\x88\x01\x01\x12\x11\n\x04page\x18\x07 \x01(\x05H\x04\x88\x01\x01\x12\x12\n\x05limit\x18\x08 \x01(\x05H\x05\x88\x01\x01\x12\x35\n\x0buser_hashes\x18\t \x01(\x0b\x32\x1b.managed_jobs.v1.UserHashesH\x06\x88\x01\x01\x12\x30\n\x08statuses\x18\n \x01(\x0b\x32\x19.managed_jobs.v1.StatusesH\x07\x88\x01\x01\x12#\n\x1bshow_jobs_without_user_hash\x18\x0b \x01(\x08\x42\n\n\x08_job_idsB\x12\n\x10_workspace_matchB\r\n\x0b_name_matchB\r\n\x0b_pool_matchB\x07\n\x05_pageB\x08\n\x06_limitB\x0e\n\x0c_user_hashesB\x0b\n\t_statuses\"\xa9\x08\n\x0eManagedJobInfo\x12\x0e\n\x06job_id\x18\x01 \x01(\x03\x12\x0f\n\x07task_id\x18\x02 \x01(\x03\x12\x10\n\x08job_name\x18\x03 \x01(\t\x12\x11\n\ttask_name\x18\x04 \x01(\t\x12\x14\n\x0cjob_duration\x18\x05 \x01(\x01\x12\x16\n\tworkspace\x18\x06 \x01(\tH\x00\x88\x01\x01\x12\x31\n\x06status\x18\x07 \x01(\x0e\x32!.managed_jobs.v1.ManagedJobStatus\x12@\n\x0eschedule_state\x18\x08 \x01(\x0e\x32(.managed_jobs.v1.ManagedJobScheduleState\x12\x11\n\tresources\x18\t \x01(\t\x12\x19\n\x11\x63luster_resources\x18\n \x01(\t\x12\x1e\n\x16\x63luster_resources_full\x18\x0b \x01(\t\x12\r\n\x05\x63loud\x18\x0c \x01(\t\x12\x0e\n\x06region\x18\r \x01(\t\x12\r\n\x05infra\x18\x0e \x01(\t\x12G\n\x0c\x61\x63\x63\x65lerators\x18\x0f \x03(\x0b\x32\x31.managed_jobs.v1.ManagedJobInfo.AcceleratorsEntry\x12\x16\n\x0erecovery_count\x18\x10 \x01(\x05\x12\x14\n\x07\x64\x65tails\x18\x11 \x01(\tH\x01\x88\x01\x01\x12\x1b\n\x0e\x66\x61ilure_reason\x18\x12 \x01(\tH\x02\x88\x01\x01\x12\x16\n\tuser_name\x18\x13 \x01(\tH\x03\x88\x01\x01\x12\x16\n\tuser_hash\x18\x14 \x01(\tH\x04\x88\x01\x01\x12\x19\n\x0csubmitted_at\x18\x15 \x01(\x01H\x05\x88\x01\x01\x12\x15\n\x08start_at\x18\x16 \x01(\x01H\x06\x88\x01\x01\x12\x13\n\x06\x65nd_at\x18\x17 \x01(\x01H\x07\x88\x01\x01\x12\x16\n\tuser_yaml\x18\x18 \x01(\tH\x08\x88\x01\x01\x12\x17\n\nentrypoint\x18\x19 \x01(\tH\t\x88\x01\x01\x12?\n\x08metadata\x18\x1a \x03(\x0b\x32-.managed_jobs.v1.ManagedJobInfo.MetadataEntry\x12\x11\n\x04pool\x18\x1b \x01(\tH\n\x88\x01\x01\x12\x16\n\tpool_hash\x18\x1c \x01(\tH\x0b\x88\x01\x01\x1a\x33\n\x11\x41\x63\x63\x65leratorsEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\r\n\x05value\x18\x02 \x01(\x02:\x02\x38\x01\x1a/\n\rMetadataEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\r\n\x05value\x18\x02 \x01(\t:\x02\x38\x01\x42\x0c\n\n_workspaceB\n\n\x08_detailsB\x11\n\x0f_failure_reasonB\x0c\n\n_user_nameB\x0c\n\n_user_hashB\x0f\n\r_submitted_atB\x0b\n\t_start_atB\t\n\x07_end_atB\x0c\n\n_user_yamlB\r\n\x0b_entrypointB\x07\n\x05_poolB\x0c\n\n_pool_hash\"\xf0\x01\n\x13GetJobTableResponse\x12-\n\x04jobs\x18\x01 \x03(\x0b\x32\x1f.managed_jobs.v1.ManagedJobInfo\x12\r\n\x05total\x18\x02 \x01(\x05\x12\x17\n\x0ftotal_no_filter\x18\x03 \x01(\x05\x12M\n\rstatus_counts\x18\x04 \x03(\x0b\x32\x36.managed_jobs.v1.GetJobTableResponse.StatusCountsEntry\x1a\x33\n\x11StatusCountsEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\r\n\x05value\x18\x02 \x01(\x05:\x02\x38\x01\"?\n\x19GetAllJobIdsByNameRequest\x12\x15\n\x08job_name\x18\x01 \x01(\tH\x00\x88\x01\x01\x42\x0b\n\t_job_name\"-\n\x1aGetAllJobIdsByNameResponse\x12\x0f\n\x07job_ids\x18\x01 \x03(\x03\"\xd7\x01\n\x11\x43\x61ncelJobsRequest\x12\x19\n\x11\x63urrent_workspace\x18\x01 \x01(\t\x12\x16\n\tuser_hash\x18\x02 \x01(\tH\x01\x88\x01\x01\x12\x13\n\tall_users\x18\x03 \x01(\x08H\x00\x12*\n\x07job_ids\x18\x04 \x01(\x0b\x32\x17.managed_jobs.v1.JobIdsH\x00\x12\x12\n\x08job_name\x18\x05 \x01(\tH\x00\x12\x13\n\tpool_name\x18\x06 \x01(\tH\x00\x42\x17\n\x15\x63\x61ncellation_criteriaB\x0c\n\n_user_hash\"%\n\x12\x43\x61ncelJobsResponse\x12\x0f\n\x07message\x18\x01 \x01(\t\"\x97\x01\n\x11StreamLogsRequest\x12\x15\n\x08job_name\x18\x01 \x01(\tH\x00\x88\x01\x01\x12\x13\n\x06job_id\x18\x02 \x01(\x03H\x01\x88\x01\x01\x12\x0e\n\x06\x66ollow\x18\x03 \x01(\x08\x12\x12\n\ncontroller\x18\x04 \x01(\x08\x12\x11\n\x04tail\x18\x05 \x01(\x05H\x02\x88\x01\x01\x42\x0b\n\t_job_nameB\t\n\x07_job_idB\x07\n\x05_tail\"L\n\x12StreamLogsResponse\x12\x10\n\x08log_line\x18\x01 \x01(\t\x12\x16\n\texit_code\x18\x02 \x01(\x05H\x00\x88\x01\x01\x42\x0c\n\n_exit_code*\x85\x04\n\x10ManagedJobStatus\x12\"\n\x1eMANAGED_JOB_STATUS_UNSPECIFIED\x10\x00\x12\x1e\n\x1aMANAGED_JOB_STATUS_PENDING\x10\x01\x12 \n\x1cMANAGED_JOB_STATUS_SUBMITTED\x10\x02\x12\x1f\n\x1bMANAGED_JOB_STATUS_STARTING\x10\x03\x12\x1e\n\x1aMANAGED_JOB_STATUS_RUNNING\x10\x04\x12!\n\x1dMANAGED_JOB_STATUS_RECOVERING\x10\x05\x12!\n\x1dMANAGED_JOB_STATUS_CANCELLING\x10\x06\x12 \n\x1cMANAGED_JOB_STATUS_SUCCEEDED\x10\x07\x12 \n\x1cMANAGED_JOB_STATUS_CANCELLED\x10\x08\x12\x1d\n\x19MANAGED_JOB_STATUS_FAILED\x10\t\x12#\n\x1fMANAGED_JOB_STATUS_FAILED_SETUP\x10\n\x12\'\n#MANAGED_JOB_STATUS_FAILED_PRECHECKS\x10\x0b\x12)\n%MANAGED_JOB_STATUS_FAILED_NO_RESOURCE\x10\x0c\x12(\n$MANAGED_JOB_STATUS_FAILED_CONTROLLER\x10\r*\x8f\x03\n\x17ManagedJobScheduleState\x12*\n&MANAGED_JOB_SCHEDULE_STATE_UNSPECIFIED\x10\x00\x12&\n\"MANAGED_JOB_SCHEDULE_STATE_INVALID\x10\x01\x12\'\n#MANAGED_JOB_SCHEDULE_STATE_INACTIVE\x10\x02\x12&\n\"MANAGED_JOB_SCHEDULE_STATE_WAITING\x10\x03\x12,\n(MANAGED_JOB_SCHEDULE_STATE_ALIVE_WAITING\x10\x04\x12(\n$MANAGED_JOB_SCHEDULE_STATE_LAUNCHING\x10\x05\x12,\n(MANAGED_JOB_SCHEDULE_STATE_ALIVE_BACKOFF\x10\x06\x12$\n MANAGED_JOB_SCHEDULE_STATE_ALIVE\x10\x07\x12#\n\x1fMANAGED_JOB_SCHEDULE_STATE_DONE\x10\x08\x32\xe4\x03\n\x12ManagedJobsService\x12U\n\nGetVersion\x12\".managed_jobs.v1.GetVersionRequest\x1a#.managed_jobs.v1.GetVersionResponse\x12X\n\x0bGetJobTable\x12#.managed_jobs.v1.GetJobTableRequest\x1a$.managed_jobs.v1.GetJobTableResponse\x12m\n\x12GetAllJobIdsByName\x12*.managed_jobs.v1.GetAllJobIdsByNameRequest\x1a+.managed_jobs.v1.GetAllJobIdsByNameResponse\x12U\n\nCancelJobs\x12\".managed_jobs.v1.CancelJobsRequest\x1a#.managed_jobs.v1.CancelJobsResponse\x12W\n\nStreamLogs\x12\".managed_jobs.v1.StreamLogsRequest\x1a#.managed_jobs.v1.StreamLogsResponse0\x01\x62\x06proto3')
+DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(b'\n*sky/schemas/generated/managed_jobsv1.proto\x12\x0fmanaged_jobs.v1\"\x15\n\x06JobIds\x12\x0b\n\x03ids\x18\x01 \x03(\x03\"\x1c\n\nUserHashes\x12\x0e\n\x06hashes\x18\x01 \x03(\t\"\x1c\n\x08Statuses\x12\x10\n\x08statuses\x18\x01 \x03(\t\"\x18\n\x06\x46ields\x12\x0e\n\x06\x66ields\x18\x01 \x03(\t\" \n\nWorkspaces\x12\x12\n\nworkspaces\x18\x01 \x03(\t\"\x13\n\x11GetVersionRequest\"0\n\x12GetVersionResponse\x12\x1a\n\x12\x63ontroller_version\x18\x01 \x01(\t\"\xe1\x04\n\x12GetJobTableRequest\x12\x15\n\rskip_finished\x18\x01 \x01(\x08\x12?\n\x15\x61\x63\x63\x65ssible_workspaces\x18\x02 \x01(\x0b\x32\x1b.managed_jobs.v1.WorkspacesH\x00\x88\x01\x01\x12-\n\x07job_ids\x18\x03 \x01(\x0b\x32\x17.managed_jobs.v1.JobIdsH\x01\x88\x01\x01\x12\x1c\n\x0fworkspace_match\x18\x04 \x01(\tH\x02\x88\x01\x01\x12\x17\n\nname_match\x18\x05 \x01(\tH\x03\x88\x01\x01\x12\x17\n\npool_match\x18\x06 \x01(\tH\x04\x88\x01\x01\x12\x11\n\x04page\x18\x07 \x01(\x05H\x05\x88\x01\x01\x12\x12\n\x05limit\x18\x08 \x01(\x05H\x06\x88\x01\x01\x12\x35\n\x0buser_hashes\x18\t \x01(\x0b\x32\x1b.managed_jobs.v1.UserHashesH\x07\x88\x01\x01\x12\x30\n\x08statuses\x18\n \x01(\x0b\x32\x19.managed_jobs.v1.StatusesH\x08\x88\x01\x01\x12#\n\x1bshow_jobs_without_user_hash\x18\x0b \x01(\x08\x12,\n\x06\x66ields\x18\x0c \x01(\x0b\x32\x17.managed_jobs.v1.FieldsH\t\x88\x01\x01\x42\x18\n\x16_accessible_workspacesB\n\n\x08_job_idsB\x12\n\x10_workspace_matchB\r\n\x0b_name_matchB\r\n\x0b_pool_matchB\x07\n\x05_pageB\x08\n\x06_limitB\x0e\n\x0c_user_hashesB\x0b\n\t_statusesB\t\n\x07_fields\"\xcb\x08\n\x0eManagedJobInfo\x12\x0e\n\x06job_id\x18\x01 \x01(\x03\x12\x0f\n\x07task_id\x18\x02 \x01(\x03\x12\x10\n\x08job_name\x18\x03 \x01(\t\x12\x11\n\ttask_name\x18\x04 \x01(\t\x12\x14\n\x0cjob_duration\x18\x05 \x01(\x01\x12\x16\n\tworkspace\x18\x06 \x01(\tH\x00\x88\x01\x01\x12\x31\n\x06status\x18\x07 \x01(\x0e\x32!.managed_jobs.v1.ManagedJobStatus\x12@\n\x0eschedule_state\x18\x08 \x01(\x0e\x32(.managed_jobs.v1.ManagedJobScheduleState\x12\x11\n\tresources\x18\t \x01(\t\x12\x19\n\x11\x63luster_resources\x18\n \x01(\t\x12\x1e\n\x16\x63luster_resources_full\x18\x0b \x01(\t\x12\r\n\x05\x63loud\x18\x0c \x01(\t\x12\x0e\n\x06region\x18\r \x01(\t\x12\r\n\x05infra\x18\x0e \x01(\t\x12G\n\x0c\x61\x63\x63\x65lerators\x18\x0f \x03(\x0b\x32\x31.managed_jobs.v1.ManagedJobInfo.AcceleratorsEntry\x12\x16\n\x0erecovery_count\x18\x10 \x01(\x05\x12\x14\n\x07\x64\x65tails\x18\x11 \x01(\tH\x01\x88\x01\x01\x12\x1b\n\x0e\x66\x61ilure_reason\x18\x12 \x01(\tH\x02\x88\x01\x01\x12\x16\n\tuser_name\x18\x13 \x01(\tH\x03\x88\x01\x01\x12\x16\n\tuser_hash\x18\x14 \x01(\tH\x04\x88\x01\x01\x12\x19\n\x0csubmitted_at\x18\x15 \x01(\x01H\x05\x88\x01\x01\x12\x15\n\x08start_at\x18\x16 \x01(\x01H\x06\x88\x01\x01\x12\x13\n\x06\x65nd_at\x18\x17 \x01(\x01H\x07\x88\x01\x01\x12\x16\n\tuser_yaml\x18\x18 \x01(\tH\x08\x88\x01\x01\x12\x17\n\nentrypoint\x18\x19 \x01(\tH\t\x88\x01\x01\x12?\n\x08metadata\x18\x1a \x03(\x0b\x32-.managed_jobs.v1.ManagedJobInfo.MetadataEntry\x12\x11\n\x04pool\x18\x1b \x01(\tH\n\x88\x01\x01\x12\x16\n\tpool_hash\x18\x1c \x01(\tH\x0b\x88\x01\x01\x12\x14\n\x07_job_id\x18\x1d \x01(\x03H\x0c\x88\x01\x01\x1a\x33\n\x11\x41\x63\x63\x65leratorsEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\r\n\x05value\x18\x02 \x01(\x02:\x02\x38\x01\x1a/\n\rMetadataEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\r\n\x05value\x18\x02 \x01(\t:\x02\x38\x01\x42\x0c\n\n_workspaceB\n\n\x08_detailsB\x11\n\x0f_failure_reasonB\x0c\n\n_user_nameB\x0c\n\n_user_hashB\x0f\n\r_submitted_atB\x0b\n\t_start_atB\t\n\x07_end_atB\x0c\n\n_user_yamlB\r\n\x0b_entrypointB\x07\n\x05_poolB\x0c\n\n_pool_hashB\n\n\x08X_job_id\"\xf0\x01\n\x13GetJobTableResponse\x12-\n\x04jobs\x18\x01 \x03(\x0b\x32\x1f.managed_jobs.v1.ManagedJobInfo\x12\r\n\x05total\x18\x02 \x01(\x05\x12\x17\n\x0ftotal_no_filter\x18\x03 \x01(\x05\x12M\n\rstatus_counts\x18\x04 \x03(\x0b\x32\x36.managed_jobs.v1.GetJobTableResponse.StatusCountsEntry\x1a\x33\n\x11StatusCountsEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\r\n\x05value\x18\x02 \x01(\x05:\x02\x38\x01\"?\n\x19GetAllJobIdsByNameRequest\x12\x15\n\x08job_name\x18\x01 \x01(\tH\x00\x88\x01\x01\x42\x0b\n\t_job_name\"-\n\x1aGetAllJobIdsByNameResponse\x12\x0f\n\x07job_ids\x18\x01 \x03(\x03\"\xd7\x01\n\x11\x43\x61ncelJobsRequest\x12\x19\n\x11\x63urrent_workspace\x18\x01 \x01(\t\x12\x16\n\tuser_hash\x18\x02 \x01(\tH\x01\x88\x01\x01\x12\x13\n\tall_users\x18\x03 \x01(\x08H\x00\x12*\n\x07job_ids\x18\x04 \x01(\x0b\x32\x17.managed_jobs.v1.JobIdsH\x00\x12\x12\n\x08job_name\x18\x05 \x01(\tH\x00\x12\x13\n\tpool_name\x18\x06 \x01(\tH\x00\x42\x17\n\x15\x63\x61ncellation_criteriaB\x0c\n\n_user_hash\"%\n\x12\x43\x61ncelJobsResponse\x12\x0f\n\x07message\x18\x01 \x01(\t\"\x97\x01\n\x11StreamLogsRequest\x12\x15\n\x08job_name\x18\x01 \x01(\tH\x00\x88\x01\x01\x12\x13\n\x06job_id\x18\x02 \x01(\x03H\x01\x88\x01\x01\x12\x0e\n\x06\x66ollow\x18\x03 \x01(\x08\x12\x12\n\ncontroller\x18\x04 \x01(\x08\x12\x11\n\x04tail\x18\x05 \x01(\x05H\x02\x88\x01\x01\x42\x0b\n\t_job_nameB\t\n\x07_job_idB\x07\n\x05_tail\"L\n\x12StreamLogsResponse\x12\x10\n\x08log_line\x18\x01 \x01(\t\x12\x16\n\texit_code\x18\x02 \x01(\x05H\x00\x88\x01\x01\x42\x0c\n\n_exit_code*\x85\x04\n\x10ManagedJobStatus\x12\"\n\x1eMANAGED_JOB_STATUS_UNSPECIFIED\x10\x00\x12\x1e\n\x1aMANAGED_JOB_STATUS_PENDING\x10\x01\x12 \n\x1cMANAGED_JOB_STATUS_SUBMITTED\x10\x02\x12\x1f\n\x1bMANAGED_JOB_STATUS_STARTING\x10\x03\x12\x1e\n\x1aMANAGED_JOB_STATUS_RUNNING\x10\x04\x12!\n\x1dMANAGED_JOB_STATUS_RECOVERING\x10\x05\x12!\n\x1dMANAGED_JOB_STATUS_CANCELLING\x10\x06\x12 \n\x1cMANAGED_JOB_STATUS_SUCCEEDED\x10\x07\x12 \n\x1cMANAGED_JOB_STATUS_CANCELLED\x10\x08\x12\x1d\n\x19MANAGED_JOB_STATUS_FAILED\x10\t\x12#\n\x1fMANAGED_JOB_STATUS_FAILED_SETUP\x10\n\x12\'\n#MANAGED_JOB_STATUS_FAILED_PRECHECKS\x10\x0b\x12)\n%MANAGED_JOB_STATUS_FAILED_NO_RESOURCE\x10\x0c\x12(\n$MANAGED_JOB_STATUS_FAILED_CONTROLLER\x10\r*\x8f\x03\n\x17ManagedJobScheduleState\x12*\n&MANAGED_JOB_SCHEDULE_STATE_UNSPECIFIED\x10\x00\x12&\n\"MANAGED_JOB_SCHEDULE_STATE_INVALID\x10\x01\x12\'\n#MANAGED_JOB_SCHEDULE_STATE_INACTIVE\x10\x02\x12&\n\"MANAGED_JOB_SCHEDULE_STATE_WAITING\x10\x03\x12,\n(MANAGED_JOB_SCHEDULE_STATE_ALIVE_WAITING\x10\x04\x12(\n$MANAGED_JOB_SCHEDULE_STATE_LAUNCHING\x10\x05\x12,\n(MANAGED_JOB_SCHEDULE_STATE_ALIVE_BACKOFF\x10\x06\x12$\n MANAGED_JOB_SCHEDULE_STATE_ALIVE\x10\x07\x12#\n\x1fMANAGED_JOB_SCHEDULE_STATE_DONE\x10\x08\x32\xe4\x03\n\x12ManagedJobsService\x12U\n\nGetVersion\x12\".managed_jobs.v1.GetVersionRequest\x1a#.managed_jobs.v1.GetVersionResponse\x12X\n\x0bGetJobTable\x12#.managed_jobs.v1.GetJobTableRequest\x1a$.managed_jobs.v1.GetJobTableResponse\x12m\n\x12GetAllJobIdsByName\x12*.managed_jobs.v1.GetAllJobIdsByNameRequest\x1a+.managed_jobs.v1.GetAllJobIdsByNameResponse\x12U\n\nCancelJobs\x12\".managed_jobs.v1.CancelJobsRequest\x1a#.managed_jobs.v1.CancelJobsResponse\x12W\n\nStreamLogs\x12\".managed_jobs.v1.StreamLogsRequest\x1a#.managed_jobs.v1.StreamLogsResponse0\x01\x62\x06proto3')
 
 _globals = globals()
 _builder.BuildMessageAndEnumDescriptors(DESCRIPTOR, _globals)
@@ -27,44 +27,48 @@ if not _descriptor._USE_C_DESCRIPTORS:
   _globals['_MANAGEDJOBINFO_METADATAENTRY']._serialized_options = b'8\001'
   _globals['_GETJOBTABLERESPONSE_STATUSCOUNTSENTRY']._loaded_options = None
   _globals['_GETJOBTABLERESPONSE_STATUSCOUNTSENTRY']._serialized_options = b'8\001'
-  _globals['_MANAGEDJOBSTATUS']._serialized_start=2625
-  _globals['_MANAGEDJOBSTATUS']._serialized_end=3142
-  _globals['_MANAGEDJOBSCHEDULESTATE']._serialized_start=3145
-  _globals['_MANAGEDJOBSCHEDULESTATE']._serialized_end=3544
+  _globals['_MANAGEDJOBSTATUS']._serialized_start=2836
+  _globals['_MANAGEDJOBSTATUS']._serialized_end=3353
+  _globals['_MANAGEDJOBSCHEDULESTATE']._serialized_start=3356
+  _globals['_MANAGEDJOBSCHEDULESTATE']._serialized_end=3755
   _globals['_JOBIDS']._serialized_start=63
   _globals['_JOBIDS']._serialized_end=84
   _globals['_USERHASHES']._serialized_start=86
   _globals['_USERHASHES']._serialized_end=114
   _globals['_STATUSES']._serialized_start=116
   _globals['_STATUSES']._serialized_end=144
-  _globals['_GETVERSIONREQUEST']._serialized_start=146
-  _globals['_GETVERSIONREQUEST']._serialized_end=165
-  _globals['_GETVERSIONRESPONSE']._serialized_start=167
-  _globals['_GETVERSIONRESPONSE']._serialized_end=215
-  _globals['_GETJOBTABLEREQUEST']._serialized_start=218
-  _globals['_GETJOBTABLEREQUEST']._serialized_end=710
-  _globals['_MANAGEDJOBINFO']._serialized_start=713
-  _globals['_MANAGEDJOBINFO']._serialized_end=1778
-  _globals['_MANAGEDJOBINFO_ACCELERATORSENTRY']._serialized_start=1512
-  _globals['_MANAGEDJOBINFO_ACCELERATORSENTRY']._serialized_end=1563
-  _globals['_MANAGEDJOBINFO_METADATAENTRY']._serialized_start=1565
-  _globals['_MANAGEDJOBINFO_METADATAENTRY']._serialized_end=1612
-  _globals['_GETJOBTABLERESPONSE']._serialized_start=1781
-  _globals['_GETJOBTABLERESPONSE']._serialized_end=2021
-  _globals['_GETJOBTABLERESPONSE_STATUSCOUNTSENTRY']._serialized_start=1970
-  _globals['_GETJOBTABLERESPONSE_STATUSCOUNTSENTRY']._serialized_end=2021
-  _globals['_GETALLJOBIDSBYNAMEREQUEST']._serialized_start=2023
-  _globals['_GETALLJOBIDSBYNAMEREQUEST']._serialized_end=2086
-  _globals['_GETALLJOBIDSBYNAMERESPONSE']._serialized_start=2088
-  _globals['_GETALLJOBIDSBYNAMERESPONSE']._serialized_end=2133
-  _globals['_CANCELJOBSREQUEST']._serialized_start=2136
-  _globals['_CANCELJOBSREQUEST']._serialized_end=2351
-  _globals['_CANCELJOBSRESPONSE']._serialized_start=2353
-  _globals['_CANCELJOBSRESPONSE']._serialized_end=2390
-  _globals['_STREAMLOGSREQUEST']._serialized_start=2393
-  _globals['_STREAMLOGSREQUEST']._serialized_end=2544
-  _globals['_STREAMLOGSRESPONSE']._serialized_start=2546
-  _globals['_STREAMLOGSRESPONSE']._serialized_end=2622
-  _globals['_MANAGEDJOBSSERVICE']._serialized_start=3547
-  _globals['_MANAGEDJOBSSERVICE']._serialized_end=4031
+  _globals['_FIELDS']._serialized_start=146
+  _globals['_FIELDS']._serialized_end=170
+  _globals['_WORKSPACES']._serialized_start=172
+  _globals['_WORKSPACES']._serialized_end=204
+  _globals['_GETVERSIONREQUEST']._serialized_start=206
+  _globals['_GETVERSIONREQUEST']._serialized_end=225
+  _globals['_GETVERSIONRESPONSE']._serialized_start=227
+  _globals['_GETVERSIONRESPONSE']._serialized_end=275
+  _globals['_GETJOBTABLEREQUEST']._serialized_start=278
+  _globals['_GETJOBTABLEREQUEST']._serialized_end=887
+  _globals['_MANAGEDJOBINFO']._serialized_start=890
+  _globals['_MANAGEDJOBINFO']._serialized_end=1989
+  _globals['_MANAGEDJOBINFO_ACCELERATORSENTRY']._serialized_start=1711
+  _globals['_MANAGEDJOBINFO_ACCELERATORSENTRY']._serialized_end=1762
+  _globals['_MANAGEDJOBINFO_METADATAENTRY']._serialized_start=1764
+  _globals['_MANAGEDJOBINFO_METADATAENTRY']._serialized_end=1811
+  _globals['_GETJOBTABLERESPONSE']._serialized_start=1992
+  _globals['_GETJOBTABLERESPONSE']._serialized_end=2232
+  _globals['_GETJOBTABLERESPONSE_STATUSCOUNTSENTRY']._serialized_start=2181
+  _globals['_GETJOBTABLERESPONSE_STATUSCOUNTSENTRY']._serialized_end=2232
+  _globals['_GETALLJOBIDSBYNAMEREQUEST']._serialized_start=2234
+  _globals['_GETALLJOBIDSBYNAMEREQUEST']._serialized_end=2297
+  _globals['_GETALLJOBIDSBYNAMERESPONSE']._serialized_start=2299
+  _globals['_GETALLJOBIDSBYNAMERESPONSE']._serialized_end=2344
+  _globals['_CANCELJOBSREQUEST']._serialized_start=2347
+  _globals['_CANCELJOBSREQUEST']._serialized_end=2562
+  _globals['_CANCELJOBSRESPONSE']._serialized_start=2564
+  _globals['_CANCELJOBSRESPONSE']._serialized_end=2601
+  _globals['_STREAMLOGSREQUEST']._serialized_start=2604
+  _globals['_STREAMLOGSREQUEST']._serialized_end=2755
+  _globals['_STREAMLOGSRESPONSE']._serialized_start=2757
+  _globals['_STREAMLOGSRESPONSE']._serialized_end=2833
+  _globals['_MANAGEDJOBSSERVICE']._serialized_start=3758
+  _globals['_MANAGEDJOBSSERVICE']._serialized_end=4242
 # @@protoc_insertion_point(module_scope)

sky/schemas/generated/managed_jobsv1_pb2.pyi

@@ -76,6 +76,18 @@ class Statuses(_message.Message):
     statuses: _containers.RepeatedScalarFieldContainer[str]
     def __init__(self, statuses: _Optional[_Iterable[str]] = ...) -> None: ...
 
+class Fields(_message.Message):
+    __slots__ = ("fields",)
+    FIELDS_FIELD_NUMBER: _ClassVar[int]
+    fields: _containers.RepeatedScalarFieldContainer[str]
+    def __init__(self, fields: _Optional[_Iterable[str]] = ...) -> None: ...
+
+class Workspaces(_message.Message):
+    __slots__ = ("workspaces",)
+    WORKSPACES_FIELD_NUMBER: _ClassVar[int]
+    workspaces: _containers.RepeatedScalarFieldContainer[str]
+    def __init__(self, workspaces: _Optional[_Iterable[str]] = ...) -> None: ...
+
 class GetVersionRequest(_message.Message):
     __slots__ = ()
     def __init__(self) -> None: ...
@@ -87,7 +99,7 @@ class GetVersionResponse(_message.Message):
     def __init__(self, controller_version: _Optional[str] = ...) -> None: ...
 
 class GetJobTableRequest(_message.Message):
-    __slots__ = ("skip_finished", "accessible_workspaces", "job_ids", "workspace_match", "name_match", "pool_match", "page", "limit", "user_hashes", "statuses", "show_jobs_without_user_hash")
+    __slots__ = ("skip_finished", "accessible_workspaces", "job_ids", "workspace_match", "name_match", "pool_match", "page", "limit", "user_hashes", "statuses", "show_jobs_without_user_hash", "fields")
     SKIP_FINISHED_FIELD_NUMBER: _ClassVar[int]
     ACCESSIBLE_WORKSPACES_FIELD_NUMBER: _ClassVar[int]
     JOB_IDS_FIELD_NUMBER: _ClassVar[int]
@@ -99,8 +111,9 @@ class GetJobTableRequest(_message.Message):
     USER_HASHES_FIELD_NUMBER: _ClassVar[int]
     STATUSES_FIELD_NUMBER: _ClassVar[int]
     SHOW_JOBS_WITHOUT_USER_HASH_FIELD_NUMBER: _ClassVar[int]
+    FIELDS_FIELD_NUMBER: _ClassVar[int]
     skip_finished: bool
-    accessible_workspaces: _containers.RepeatedScalarFieldContainer[str]
+    accessible_workspaces: Workspaces
     job_ids: JobIds
     workspace_match: str
     name_match: str
@@ -110,10 +123,11 @@ class GetJobTableRequest(_message.Message):
     user_hashes: UserHashes
     statuses: Statuses
     show_jobs_without_user_hash: bool
-    def __init__(self, skip_finished: bool = ..., accessible_workspaces: _Optional[_Iterable[str]] = ..., job_ids: _Optional[_Union[JobIds, _Mapping]] = ..., workspace_match: _Optional[str] = ..., name_match: _Optional[str] = ..., pool_match: _Optional[str] = ..., page: _Optional[int] = ..., limit: _Optional[int] = ..., user_hashes: _Optional[_Union[UserHashes, _Mapping]] = ..., statuses: _Optional[_Union[Statuses, _Mapping]] = ..., show_jobs_without_user_hash: bool = ...) -> None: ...
+    fields: Fields
+    def __init__(self, skip_finished: bool = ..., accessible_workspaces: _Optional[_Union[Workspaces, _Mapping]] = ..., job_ids: _Optional[_Union[JobIds, _Mapping]] = ..., workspace_match: _Optional[str] = ..., name_match: _Optional[str] = ..., pool_match: _Optional[str] = ..., page: _Optional[int] = ..., limit: _Optional[int] = ..., user_hashes: _Optional[_Union[UserHashes, _Mapping]] = ..., statuses: _Optional[_Union[Statuses, _Mapping]] = ..., show_jobs_without_user_hash: bool = ..., fields: _Optional[_Union[Fields, _Mapping]] = ...) -> None: ...
 
 class ManagedJobInfo(_message.Message):
-    __slots__ = ("job_id", "task_id", "job_name", "task_name", "job_duration", "workspace", "status", "schedule_state", "resources", "cluster_resources", "cluster_resources_full", "cloud", "region", "infra", "accelerators", "recovery_count", "details", "failure_reason", "user_name", "user_hash", "submitted_at", "start_at", "end_at", "user_yaml", "entrypoint", "metadata", "pool", "pool_hash")
+    __slots__ = ("job_id", "task_id", "job_name", "task_name", "job_duration", "workspace", "status", "schedule_state", "resources", "cluster_resources", "cluster_resources_full", "cloud", "region", "infra", "accelerators", "recovery_count", "details", "failure_reason", "user_name", "user_hash", "submitted_at", "start_at", "end_at", "user_yaml", "entrypoint", "metadata", "pool", "pool_hash", "_job_id")
     class AcceleratorsEntry(_message.Message):
         __slots__ = ("key", "value")
         KEY_FIELD_NUMBER: _ClassVar[int]
@@ -156,6 +170,7 @@ class ManagedJobInfo(_message.Message):
     METADATA_FIELD_NUMBER: _ClassVar[int]
     POOL_FIELD_NUMBER: _ClassVar[int]
     POOL_HASH_FIELD_NUMBER: _ClassVar[int]
+    _JOB_ID_FIELD_NUMBER: _ClassVar[int]
     job_id: int
     task_id: int
     job_name: str
@@ -184,7 +199,8 @@ class ManagedJobInfo(_message.Message):
     metadata: _containers.ScalarMap[str, str]
     pool: str
     pool_hash: str
-    def __init__(self, job_id: _Optional[int] = ..., task_id: _Optional[int] = ..., job_name: _Optional[str] = ..., task_name: _Optional[str] = ..., job_duration: _Optional[float] = ..., workspace: _Optional[str] = ..., status: _Optional[_Union[ManagedJobStatus, str]] = ..., schedule_state: _Optional[_Union[ManagedJobScheduleState, str]] = ..., resources: _Optional[str] = ..., cluster_resources: _Optional[str] = ..., cluster_resources_full: _Optional[str] = ..., cloud: _Optional[str] = ..., region: _Optional[str] = ..., infra: _Optional[str] = ..., accelerators: _Optional[_Mapping[str, float]] = ..., recovery_count: _Optional[int] = ..., details: _Optional[str] = ..., failure_reason: _Optional[str] = ..., user_name: _Optional[str] = ..., user_hash: _Optional[str] = ..., submitted_at: _Optional[float] = ..., start_at: _Optional[float] = ..., end_at: _Optional[float] = ..., user_yaml: _Optional[str] = ..., entrypoint: _Optional[str] = ..., metadata: _Optional[_Mapping[str, str]] = ..., pool: _Optional[str] = ..., pool_hash: _Optional[str] = ...) -> None: ...
+    _job_id: int
+    def __init__(self, job_id: _Optional[int] = ..., task_id: _Optional[int] = ..., job_name: _Optional[str] = ..., task_name: _Optional[str] = ..., job_duration: _Optional[float] = ..., workspace: _Optional[str] = ..., status: _Optional[_Union[ManagedJobStatus, str]] = ..., schedule_state: _Optional[_Union[ManagedJobScheduleState, str]] = ..., resources: _Optional[str] = ..., cluster_resources: _Optional[str] = ..., cluster_resources_full: _Optional[str] = ..., cloud: _Optional[str] = ..., region: _Optional[str] = ..., infra: _Optional[str] = ..., accelerators: _Optional[_Mapping[str, float]] = ..., recovery_count: _Optional[int] = ..., details: _Optional[str] = ..., failure_reason: _Optional[str] = ..., user_name: _Optional[str] = ..., user_hash: _Optional[str] = ..., submitted_at: _Optional[float] = ..., start_at: _Optional[float] = ..., end_at: _Optional[float] = ..., user_yaml: _Optional[str] = ..., entrypoint: _Optional[str] = ..., metadata: _Optional[_Mapping[str, str]] = ..., pool: _Optional[str] = ..., pool_hash: _Optional[str] = ..., _job_id: _Optional[int] = ...) -> None: ...
 
 class GetJobTableResponse(_message.Message):
     __slots__ = ("jobs", "total", "total_no_filter", "status_counts")

sky/serve/replica_managers.py

@@ -495,8 +495,8 @@ class ReplicaInfo:
                 info_dict['cloud'] = repr(handle.launched_resources.cloud)
                 info_dict['region'] = handle.launched_resources.region
                 info_dict['resources_str'] = (
-                    resources_utils.get_readable_resources_repr(handle,
-                                                                simplify=True))
+                    resources_utils.get_readable_resources_repr(
+                        handle, simplified_only=True)[0])
         return info_dict
 
     def __repr__(self) -> str:

sky/serve/serve_utils.py

@@ -1550,8 +1550,15 @@ def _format_replica_table(replica_records: List[Dict[str, Any]], show_all: bool,
             'handle']
         if replica_handle is not None:
             infra = replica_handle.launched_resources.infra.formatted_str()
-            resources_str = resources_utils.get_readable_resources_repr(
-                replica_handle, simplify=not show_all)
+            simplified = not show_all
+            resources_str_simple, resources_str_full = (
+                resources_utils.get_readable_resources_repr(
+                    replica_handle, simplified_only=simplified))
+            if simplified:
+                resources_str = resources_str_simple
+            else:
+                assert resources_str_full is not None
+                resources_str = resources_str_full
 
         replica_values = [
             service_name,

sky/serve/server/server.py

@@ -10,6 +10,7 @@ from sky.server import common as server_common
 from sky.server import stream_utils
 from sky.server.requests import executor
 from sky.server.requests import payloads
+from sky.server.requests import request_names
 from sky.server.requests import requests as api_requests
 from sky.skylet import constants
 from sky.utils import common
@@ -25,7 +26,7 @@ async def up(
 ) -> None:
     await executor.schedule_request_async(
         request_id=request.state.request_id,
-        request_name='serve.up',
+        request_name=request_names.RequestName.SERVE_UP,
         request_body=up_body,
         func=core.up,
         schedule_type=api_requests.ScheduleType.LONG,
@@ -40,7 +41,7 @@ async def update(
 ) -> None:
     await executor.schedule_request_async(
         request_id=request.state.request_id,
-        request_name='serve.update',
+        request_name=request_names.RequestName.SERVE_UPDATE,
         request_body=update_body,
         func=core.update,
         schedule_type=api_requests.ScheduleType.SHORT,
@@ -55,7 +56,7 @@ async def down(
 ) -> None:
     await executor.schedule_request_async(
         request_id=request.state.request_id,
-        request_name='serve.down',
+        request_name=request_names.RequestName.SERVE_DOWN,
         request_body=down_body,
         func=core.down,
         schedule_type=api_requests.ScheduleType.SHORT,
@@ -70,7 +71,7 @@ async def terminate_replica(
 ) -> None:
     await executor.schedule_request_async(
         request_id=request.state.request_id,
-        request_name='serve.terminate_replica',
+        request_name=request_names.RequestName.SERVE_TERMINATE_REPLICA,
         request_body=terminate_replica_body,
         func=core.terminate_replica,
         schedule_type=api_requests.ScheduleType.SHORT,
@@ -85,7 +86,7 @@ async def status(
 ) -> None:
     await executor.schedule_request_async(
         request_id=request.state.request_id,
-        request_name='serve.status',
+        request_name=request_names.RequestName.SERVE_STATUS,
         request_body=status_body,
         func=core.status,
         schedule_type=api_requests.ScheduleType.SHORT,
@@ -101,7 +102,7 @@ async def tail_logs(
     executor.check_request_thread_executor_available()
     request_task = await executor.prepare_request_async(
         request_id=request.state.request_id,
-        request_name='serve.logs',
+        request_name=request_names.RequestName.SERVE_LOGS,
         request_body=log_body,
         func=core.tail_logs,
         schedule_type=api_requests.ScheduleType.SHORT,
@@ -134,7 +135,7 @@ async def download_logs(
     download_logs_body.local_dir = str(logs_dir_on_api_server)
     await executor.schedule_request_async(
         request_id=request.state.request_id,
-        request_name='serve.sync_down_logs',
+        request_name=request_names.RequestName.SERVE_SYNC_DOWN_LOGS,
         request_body=download_logs_body,
         func=core.sync_down_logs,
         schedule_type=api_requests.ScheduleType.SHORT,