Merge branch 'feature/ET-684/run-centric-api' into aaron_amanuel/sear…

…ch-actions

.bumpversion.cfg

@@ -1,5 +1,5 @@
 [bumpversion]
-current_version = 0.36.1-dev0
+current_version = 0.37.1-dev0
 commit = true
 tag = true
 tag_name = {new_version}

.circleci/real_config.yml

@@ -21,7 +21,7 @@ executors:
 parameters:
   det-version:
     type: string
-    default: 0.36.1-dev0
+    default: 0.37.1-dev0
   docker-image:
     type: string
     default: determinedai/cimg-base:latest

VERSION

@@ -1 +1 @@
-0.36.1-dev0
+0.37.1-dev0

docs/.redirects/all_published_urls_ever.json

@@ -127,6 +127,7 @@
     "integrations/notification/_index",
     "integrations/notification/index",
     "integrations/notification/slack",
+    "integrations/notification/workload-alerting",
     "integrations/notification/zapier",
     "integrations/observability/_index",
     "integrations/pachyderm/pachyderm",
@@ -148,6 +149,7 @@
     "introduction",
     "join-community",
     "manage/_index",
+    "manage/cluster-overview",
     "manage/elasticsearch-logging-backend",
     "manage/historical-cluster-usage-data",
     "manage/security/_index",
@@ -434,6 +436,7 @@
     "setup-cluster/on-prem/options/linux-packages",
     "setup-cluster/on-prem/options/wsl",
     "setup-cluster/on-prem/requirements",
+    "setup-cluster/rocm-support",
     "setup-cluster/security/index",
     "setup-cluster/security/oauth",
     "setup-cluster/security/oidc",
@@ -444,6 +447,7 @@
     "setup-cluster/security/tls",
     "setup-cluster/setup-clients",
     "setup-cluster/slurm/_index",
+    "setup-cluster/slurm/hpc-environment-requirements",
     "setup-cluster/slurm/hpc-launching-architecture",
     "setup-cluster/slurm/hpc-security-considerations",
     "setup-cluster/slurm/hpc-with-agent",

docs/_static/version-switcher/versions.json

@@ -1,8 +1,12 @@
 [
     {
-        "version": "0.36.1-dev0",
+        "version": "0.37.1-dev0",
         "url": "https://docs.determined.ai/latest/"
     },
+    {
+        "version": "0.37.0",
+        "url": "https://docs.determined.ai/0.37.0/"
+    },
     {
         "version": "0.36.0",
         "url": "https://docs.determined.ai/0.36.0/"

docs/deploy/upload.py

@@ -3,6 +3,7 @@
 import mimetypes
 import os
 import pathlib
+import time
 
 import boto3
 

docs/model-dev-guide/hyperparameter/search-methods/_index.rst

@@ -38,8 +38,8 @@ checkpoints for downstream serving.
  Adaptive Search
 *****************
 
-Our default recommended search method is `Adaptive (ASHA) <https://arxiv.org/pdf/1810.05934.pdf>`_,
-a state-of-the-art early-stopping based technique that speeds up traditional techniques like random
+Our default recommended search method is `Adaptive (ASHA) <http://arxiv.org/pdf/1810.05934>`_, a
+state-of-the-art early-stopping based technique that speeds up traditional techniques like random
 search by periodically abandoning low-performing hyperparameter configurations in a principled
 fashion.
 

docs/reference/deploy/helm-config-reference.rst

@@ -77,6 +77,25 @@
       PersistentVolumes is disabled, users must manually create a PersistentVolume that will match
       the PersistentVolumeClaim.
 
+   -  ``claimSuffix``: Optional configuration that defines the persistent volume claim name for the
+      Determined database. If not undefined, a new PVC is created, with the name
+      ``determined-db-pvc-$releaseName``. If specified, the provided value will be used as the
+      suffix to ``determined-db-pvc-``.
+
+   -  ``snapshotSuffix``: Optional configuration for naming the volume snapshot, which serves as a
+      backup of the Determined database's persistentVolume. If defined, Helm will create a snapshot
+      of the database during the next upgrade, creating a volumeSnapshot named
+      ``determined-db-snapsnot-$snapshotSuffix``. This **must** **not** match the name of an
+      existing persistent volume claim, such as ``determined-db-pvc-$releaseName``, because the
+      ``snapshotSuffix`` will be used to create a new persistent volume claim when restoring.
+
+   -  ``restoreSnapshotSuffix``: Optional configuration of the volumeSnapshot to restore from during
+      an upgrade. If this is set during an upgrade, a new persistentVolume will be created and
+      swapped into Determined's database by creating a new persistent volume claim named
+      ``determined-db-pvc-$restoreSnapshotSuffix``, restoring data from
+      ``determined-db-snapsnot-$restoreSnapshotSuffix``. This **must not** match the name of an
+      existing persistent volume claim (PVC), such as ``$releaseName``.
+
 -  ``checkpointStorage``: Specifies where model checkpoints will be stored. This can be overridden
    on a per-experiment basis in the :ref:`experiment-config-reference`. A checkpoint contains the
    architecture and weights of the model being trained. Determined currently supports several kinds

docs/reference/deploy/master-config-reference.rst

@@ -133,8 +133,8 @@ Defaults to ``false``.
 =================
 
 Defines the default `Docker registry credentials
-<https://docs.docker.com/engine/api/v1.30/#tag/System/operation/SystemAuth>`__ to use when pulling a
-custom base Docker image, if needed. If credentials are specified in the :ref:`experiment config
+<https://docs.docker.com/reference/api/engine/version/v1.30/>`__ to use when pulling a custom base
+Docker image, if needed. If credentials are specified in the :ref:`experiment config
 <exp-environment-image>` this default value is overridden. Credentials are specified as the
 following nested fields:
 
@@ -283,7 +283,7 @@ master config that provides a readable name for the Determined deployment.
 ``name``
 --------
 
-(deprecated) Specifies the resource manager’s name. ``cluster_name`` should be specified instead.
+(deprecated) Specifies the resource manager's name. ``cluster_name`` should be specified instead.
 
 ``metadata``
 ============

docs/reference/experiment-config-reference.rst

@@ -1313,8 +1313,8 @@ Optional. Whether the bind-mount should be a read-only mount. Defaults to ``fals
 ===============
 
 Optional. `Propagation behavior
-<https://docs.docker.com/storage/bind-mounts/#configure-bind-propagation>`__ for replicas of the
-bind-mount. Defaults to ``rprivate``.
+<https://docs.docker.com/engine/storage/bind-mounts/#configure-bind-propagation>`__ for replicas of
+the bind-mount. Defaults to ``rprivate``.
 
 For example, to mount ``/data`` on the host to the same path in the container, use:
 

docs/release-notes.rst

@@ -6,6 +6,52 @@
  Release Notes
 ###############
 
+**************
+ Version 0.37
+**************
+
+Version 0.37.0
+==============
+
+**Release Date:** September 27, 2024
+
+**Breaking Changes**
+
+-  API: Remove the ``model_hub`` library from Determined.
+
+-  Starting with this release, ``MMDetTrial`` and ``BaseTransformerTrial`` are removed. HuggingFace
+   users should refer to the provided `HuggingFace TrainerAPI examples
+   <https://github.com/determined-ai/determined/tree/main/examples/hf_trainer_api>`__, which use a
+   custom callback instead of BaseTransformerTrial. Users of ``MMDetTrial`` can refer to :ref:`Core
+   API <api-core-ug>`.
+
+**New Features**
+
+-  Webhooks: Add support for experiment monitoring and alerting. Capabilities include
+   workspace-level subscriptions for "All experiments" or "Specific experiment(s) with matching
+   configuration" options. New trigger types include ``COMPLETED``, ``ERROR``, ``TASKLOG``, and
+   ``CUSTOM``. Support for custom triggers, code-based alerts, experiment-specific webhook
+   exclusions, and editable webhook URLs is also added. For details, visit
+   :ref:`supported-webhook-triggers`.
+
+-  Master Configuration: Add support for POSIX claims in the master configuration. It now accepts
+   ``agent_uid_attribute_name``, ``agent_gid_attribute_name``, ``agent_user_name_attribute_name``,
+   or ``agent_group_name_attribute_name``. Refer to the :ref:`OIDC master configuration
+   <master-config-oidc>` or :ref:`SAML master configuration <master-config-saml>` for details. If
+   any of these fields are configured, they will sync with the database.
+
+**Improvements**
+
+-  WebUI: Change the "Compute Slots Allocated" label to "Unspecified Slots Allocated" for resource
+   pools with no or multiple slot types. Add error logs for zero or multi-slot-type cases and update
+   the progress bar to include all agents when the slot type is ``TYPE_UNSPECIFIED``.
+
+**Bug Fixes**
+
+-  API/Tasks: Fix a bug where a master-configured ``log_retention_days`` value is not applied to
+   experiments and tasks. The master-configured value is now correctly applied to new experiments,
+   and all pre-existing experiments will also follow the specified ``log_retention_days``.
+
 **************
  Version 0.36
 **************

docs/release-notes/9966-fix-grid.rst

@@ -0,0 +1,7 @@
+:orphan:
+
+**Fixes**
+
+-  Previously, during a grid search, if a hyperparameter contained an empty nested hyperparameter
+   (that is, just an empty map), that hyperparameter would not appear in the hparams passed to the
+   trial.

docs/release-notes/helm-db-snapshot.rst

@@ -0,0 +1,6 @@
+:orphan:
+
+**New Features**
+
+-  Helm: Add support for capturing and restoring snapshots of the database persistent volume. Visit
+   :ref:`helm-config-reference` for more details.

docs/setup-cluster/on-prem/options/docker.rst

@@ -193,7 +193,8 @@ large range of ports, as it does not require network address translation (NAT) a
 The host networking driver only works on Linux hosts, and is not supported on Docker Desktop for
 Mac, Docker Desktop for Windows, or Docker EE for Windows Server.
 
-See `Docker's documentation <https://docs.docker.com/network/drivers/host/>`_ for more details.
+See `Docker's documentation <https://docs.docker.com/engine/network/drivers/host/>`_ for more
+details.
 
 .. note::
 
@@ -209,7 +210,7 @@ See `Docker's documentation <https://docs.docker.com/network/drivers/host/>`_ fo
 
 By default, ``docker run`` will run in the foreground, so that a container can be stopped simply by
 pressing Control-C. If you wish to keep Determined running for the long term, consider running the
-containers `detached <ttps://docs.docker.com/engine/reference/commandline/container_run/#detach>`_
+containers `detached <https://docs.docker.com/engine/reference/commandline/container_run/#detach>`_
 and/or with `restart policies
 <https://docs.docker.com/engine/containers/start-containers-automatically/>`_. Using :ref:`our
 deployment tool <install-using-deploy>` is also an option.

e2e_tests/tests/cluster/test_priority_scheduler.py

@@ -34,15 +34,15 @@ def test_priortity_scheduler_noop_command(
     managed_cluster_priority_scheduler.ensure_agent_ok()
     assert str(conf.MASTER_PORT) == "8082"
     # without slots (and default priority)
-    command_id = utils.run_command(sess, slots=0)
+    command_id = utils.run_command(sess, 0, slots=0)
     utils.wait_for_command_state(sess, command_id, "TERMINATED", 40)
     utils.assert_command_succeeded(sess, command_id)
     # with slots (and default priority)
-    command_id = utils.run_command(sess, slots=1)
+    command_id = utils.run_command(sess, 0, slots=1)
     utils.wait_for_command_state(sess, command_id, "TERMINATED", 60)
     utils.assert_command_succeeded(sess, command_id)
     # explicity priority
-    command_id = utils.run_command_set_priority(sess, slots=0, priority=60)
+    command_id = utils.run_command_set_priority(sess, 0, slots=0, priority=60)
     utils.wait_for_command_state(sess, command_id, "TERMINATED", 60)
     utils.assert_command_succeeded(sess, command_id)
 

e2e_tests/tests/cluster/utils.py

@@ -101,7 +101,7 @@ def num_free_slots(sess: api.Session) -> int:
 
 
 def run_command_set_priority(
-    sess: api.Session, sleep: int = 30, slots: int = 1, priority: int = 0
+    sess: api.Session, sleep: int, slots: int = 1, priority: int = 0
 ) -> str:
     cmd = [
         "det",
@@ -118,7 +118,7 @@ def run_command_set_priority(
     return detproc.check_output(sess, cmd).strip()
 
 
-def run_command(sess: api.Session, sleep: int = 30, slots: int = 1) -> str:
+def run_command(sess: api.Session, sleep: int, slots: int = 1) -> str:
     cmd = [
         "det",
         "command",
@@ -144,7 +144,7 @@ def run_command_args(sess: api.Session, entrypoint: str, args: Optional[List[str
     return detproc.check_output(sess, cmd + [entrypoint]).strip()
 
 
-def run_zero_slot_command(sess: api.Session, sleep: int = 30) -> str:
+def run_zero_slot_command(sess: api.Session, sleep: int) -> str:
     return run_command(sess, sleep=sleep, slots=0)
 
 

harness/determined/__version__.py

@@ -1 +1 @@
-__version__ = "0.36.1-dev0"
+__version__ = "0.37.1-dev0"