[DPE-6955] Integrate Advanced Statuses in MongoDB K8S (#402)

Gu1nness · web-flow · commit 65f70b7cee26 · 2025-06-24T17:13:23.000+02:00
* wip: statuses integration

* fix: tls + charmcraft

* fix: tests

* fix: relock

* fix: more fixes

* fix: post rebase

* fix: unit tests

* fix: charm version

* fix: relock

* fix: relock

* fix: relock

* fix: extend ci

* revert: extend ci

* fix: remove constraints after update of pydantic

* fix: relock + fix test

* fix: relock

* fix: relock

* fix: don't remove shard on scale down

* fix: relock + unit test fixes

* fix: relock

* fix: relock

* fix: poetry

* feat: use published library

* fix: remove breakpoint

* fix: use status-detail to get full message

* feat: Add CLA checker
diff --git a/.github/workflows/cla.yaml b/.github/workflows/cla.yaml
@@ -0,0 +1,9 @@
+name: cla-check
+on: [pull_request]
+
+jobs:
+  cla-check:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Check if CLA signed
+        uses: canonical/has-signed-canonical-cla@v2
diff --git a/actions.yaml b/actions.yaml
@@ -71,3 +71,11 @@ force-refresh-start:
   description: |
     Force refresh of this unit.
     Potential of data loss and downtime.
+
+status-detail:
+  description: Gets statuses of the charm
+  params:
+    recompute:
+      type: boolean
+      default: false
+      description: a boolean indicating whether a unit should recompute all statuses.
diff --git a/charmcraft.yaml b/charmcraft.yaml
@@ -51,8 +51,6 @@ parts:
     after:
       - poetry-deps
     poetry-export-extra-args: ['--only', 'main,charm-libs']
-    build-environment:
-      - PIP_CONSTRAINT: constraints.txt # Workaround for typing-extensions / pydantic-core incompatibility
     build-packages:
       - libffi-dev  # Needed to build Python dependencies with Rust from source
       - libssl-dev  # Needed to build Python dependencies with Rust from source
diff --git a/constraints.txt b/constraints.txt
diff --git a/metadata.yaml b/metadata.yaml
@@ -26,6 +26,8 @@ peers:
   ldap-peers:
     # Stores the data for the ldap relation.
     interface: ldap-peers
+  status-peers:
+    interface: status-peers
 
 provides:
   database:
diff --git a/poetry.lock b/poetry.lock
diff --git a/pyproject.toml b/pyproject.toml
@@ -8,7 +8,7 @@ requires-poetry = ">=2.0.0"
 
 [tool.poetry.dependencies]
 python = "^3.10.12"
-mongo-charms-single-kernel = "==0.0.6"
+mongo-charms-single-kernel = "0.0.9"
 ops = "~2.15.0"
 pymongo = "^4.7.3"
 tenacity = "^8.2.3"
@@ -20,7 +20,6 @@ overrides = "^7.7.0"
 lightkube = "^0.15.3"
 setuptools = "^72.0.0"
 rpds-py = "0.18.0"
-typing-extensions = "<4.14" # Workaround for typing-extensions / pydantic-core incompatibility
 
 [tool.poetry.group.charm-libs.dependencies]
 ops = "~2.15.0"
@@ -58,12 +57,12 @@ coverage = {extras = ["toml"], version = "^7.5.0"}
 pytest = "^8.1.1"
 pytest-mock = "*"
 parameterized = "^0.9.0"
-mongo-charms-single-kernel = "==0.0.6"
+mongo-charms-single-kernel = "0.0.9"
 
 [tool.poetry.group.integration.dependencies]
 allure-pytest = "^2.13.5"
 ops = "~2.15.0"
-mongo-charms-single-kernel = "==0.0.6"
+mongo-charms-single-kernel = "0.0.9"
 tenacity = "^8.2.3"
 pymongo = "^4.7.3"
 parameterized = "^0.9.0"
diff --git a/tests/integration/helpers.py b/tests/integration/helpers.py
@@ -110,7 +110,10 @@ async def get_address_of_unit(ops_test: OpsTest, unit_id: int, app_name: str = A
 
 
 async def get_password(
-    ops_test: OpsTest, unit_id: int = 0, username: str = "operator", app_name: str = APP_NAME
+    ops_test: OpsTest,
+    unit_id: int = 0,
+    username: str = "operator",
+    app_name: str = APP_NAME,
 ) -> str:
     """Use the charm action to retrieve the password from provided unit.
 
@@ -670,8 +673,9 @@ async def check_all_units_blocked_with_status(
             unit.workload_status.value == "blocked"
         ), f"unit {unit.name} not in blocked state, in {unit.workload_status.value}"
         if status:
+            # We can have extra info but we care for the most important status
             assert (
-                unit.workload_status.message == status
+                status in unit.workload_status.message
             ), f"unit {unit.name} not in blocked state, in {unit.workload_status.value}"
 
 
@@ -693,6 +697,20 @@ async def wait_for_mongodb_units_blocked(
         await ops_test.model.set_config({hook_interval_key: old_interval})
 
 
+async def check_status_detail(ops_test: OpsTest, app_name: str, status: str, message: str) -> None:
+    """Checks that the first status returned by status-detail is the one expected."""
+    for unit in ops_test.model.applications[app_name].units:
+        action = await unit.run_action("status-detail")
+        action = await action.wait()
+        result = action.results["json-output"]
+
+        # juju messes up the string formatting here.
+        unit_statuses = json.loads(result["unit"].replace("'", '"'))
+
+        assert unit_statuses[0]["Status"].lower() == status
+        assert unit_statuses[0]["Message"] == message
+
+
 def is_relation_joined(ops_test: OpsTest, endpoint_one: str, endpoint_two: str) -> bool:
     """Check if a relation is joined.
 
diff --git a/tests/integration/sharding_tests/test_sharding.py b/tests/integration/sharding_tests/test_sharding.py
@@ -37,7 +37,7 @@
 BACKUP_USERNAME = "backup"
 PASSWORD = "operator-password"
 DIFFERENT_PASSWORD = "shard-set-password"
-CONFIG_SERVER_NEEDS_SHARD_STATUS = "missing relation to shard(s)"
+CONFIG_SERVER_NEEDS_SHARD_STATUS = "Missing relation to shard(s)."
 SHARD_NEEDS_CONFIG_SERVER_STATUS = "Missing relation to config-server."
 
 # for now we have a large timeout due to the slow drainage of the `config.system.sessions`
@@ -220,6 +220,7 @@ async def test_sharding(ops_test: OpsTest) -> None:
 
 @pytest.mark.group(1)
 @pytest.mark.parametrize("username", [OPERATOR_USERNAME, BACKUP_USERNAME])
+@pytest.mark.abort_on_fail
 async def test_set_operator_password(ops_test: OpsTest, username):
     """Tests that the cluster can safely set the operator password."""
     for cluster_app_name in CLUSTER_APPS:
@@ -273,6 +274,7 @@ async def test_set_operator_password(ops_test: OpsTest, username):
 
 
 @pytest.mark.group(1)
+@pytest.mark.abort_on_fail
 async def test_shard_removal(ops_test: OpsTest) -> None:
     """Test shard removal.
 
@@ -329,6 +331,7 @@ async def test_shard_removal(ops_test: OpsTest) -> None:
 
 
 @pytest.mark.group(1)
+@pytest.mark.abort_on_fail
 async def test_removal_of_non_primary_shard(ops_test: OpsTest):
     """Tests safe removal of a shard that is not primary."""
     # add back a shard so we can safely remove a shard.
@@ -381,6 +384,7 @@ async def test_removal_of_non_primary_shard(ops_test: OpsTest):
 
 
 @pytest.mark.group(1)
+@pytest.mark.abort_on_fail
 async def test_unconventual_shard_removal(ops_test: OpsTest):
     """Tests that removing a shard application safely drains data.
 
diff --git a/tests/integration/sharding_tests/test_sharding_relations.py b/tests/integration/sharding_tests/test_sharding_relations.py
@@ -5,7 +5,12 @@
 from juju.errors import JujuAPIError
 from pytest_operator.plugin import OpsTest
 
-from ..helpers import DEPLOYMENT_TIMEOUT, METADATA, wait_for_mongodb_units_blocked
+from ..helpers import (
+    DEPLOYMENT_TIMEOUT,
+    METADATA,
+    check_status_detail,
+    wait_for_mongodb_units_blocked,
+)
 
 S3_APP_NAME = "s3-integrator"
 SHARD_ONE_APP_NAME = "shard"
@@ -163,7 +168,7 @@ async def test_replication_config_server_relation(ops_test: OpsTest):
     await wait_for_mongodb_units_blocked(
         ops_test,
         REPLICATION_APP_NAME,
-        status="sharding interface cannot be used by replicas",
+        status="Sharding interface cannot be used by replicas.",
         timeout=300,
     )
 
@@ -193,7 +198,7 @@ async def test_replication_shard_relation(ops_test: OpsTest):
     await wait_for_mongodb_units_blocked(
         ops_test,
         REPLICATION_APP_NAME,
-        status="sharding interface cannot be used by replicas",
+        status="Sharding interface cannot be used by replicas.",
         timeout=300,
     )
 
@@ -255,9 +260,15 @@ async def test_shard_mongos_relation(ops_test: OpsTest) -> None:
     await wait_for_mongodb_units_blocked(
         ops_test,
         SHARD_ONE_APP_NAME,
-        status="Relation to mongos not supported, config role must be config-server",
+        status="Relation to mongos not supported",
         timeout=300,
     )
+    await check_status_detail(
+        ops_test,
+        SHARD_ONE_APP_NAME,
+        status="blocked",
+        message="Relation to mongos not supported, config role must be config-server.",
+    )
 
     # clean up relations
     await ops_test.model.applications[SHARD_ONE_APP_NAME].remove_relation(
@@ -285,9 +296,15 @@ async def test_shard_s3_relation(ops_test: OpsTest) -> None:
     await wait_for_mongodb_units_blocked(
         ops_test,
         SHARD_ONE_APP_NAME,
-        status="Relation to s3-integrator is not supported, config role must be config-server.",
+        status="Relation to s3-integrator is not support",
         timeout=300,
     )
+    await check_status_detail(
+        ops_test,
+        SHARD_ONE_APP_NAME,
+        status="blocked",
+        message="Relation to s3-integrator is not supported, config role must be config-server.",
+    )
 
     # clean up relations
     await ops_test.model.applications[SHARD_ONE_APP_NAME].remove_relation(
diff --git a/tests/integration/upgrades/test_revision_check.py b/tests/integration/upgrades/test_revision_check.py
@@ -96,10 +96,6 @@ async def test_local_shard_reports_remote_config_server(ops_test: OpsTest) -> No
     await wait_for_mongodb_units_blocked(
         ops_test,
         LOCAL_SHARD_APP_NAME,
-        timeout=300,
+        timeout=600,
+        status="is not up-to date with config-server",
     )
-
-    shard_unit = ops_test.model.applications[LOCAL_SHARD_APP_NAME].units[0]
-    assert (
-        "is not up-to date with config-server" in shard_unit.workload_status_message
-    ), "Shard does not correctly report mismatch in revision"
diff --git a/tests/unit/test_charm.py b/tests/unit/test_charm.py
@@ -6,7 +6,7 @@
 from unittest.mock import patch
 
 import pytest
-from ops.model import MaintenanceStatus
+from data_platform_helpers.advanced_statuses.models import StatusObject
 from ops.pebble import PathError, ProtocolError
 from ops.testing import ActionFailed, Harness
 from parameterized import parameterized
@@ -650,12 +650,11 @@ def test_delete_password(self):
             )
 
     @parameterized.expand([(Scope.APP), (Scope.UNIT)])
-    @patch("single_kernel_mongo.status.StatusManager.process_and_share_statuses")
     @patch("single_kernel_mongo.managers.config.BackupConfigManager.configure_and_restart")
     @patch(
         "single_kernel_mongo.managers.config.MongoDBExporterConfigManager.configure_and_restart"
     )
-    def test_on_secret_changed(self, scope, connect_exporter, connect_backup, *unused):
+    def test_on_secret_changed(self, scope, connect_exporter, connect_backup):
         """NOTE: currently ops.testing seems to allow for non-leader to set secrets too!"""
         secret = self.harness.charm.operator.state.secrets.set("new-secret", "bla", scope)
         secret = self.harness.charm.model.get_secret(label=secret.label)
@@ -790,12 +789,12 @@ def test_set_password_provided(self, *unused):
         # verify app data is updated and results are reported to user
         self.assertEqual("canonical123", new_password)
 
-    @patch("single_kernel_mongo.managers.backups.BackupManager.get_status")
+    @patch("single_kernel_mongo.managers.backups.BackupManager.get_statuses")
     def test_set_backup_password_pbm_busy(self, pbm_status):
         """Tests changes to passwords fail when pbm is restoring/backing up."""
         self.harness.set_leader(True)
 
-        pbm_status.return_value = MaintenanceStatus("pbm")
+        pbm_status.return_value = [StatusObject(status="maintenance", message="pbm")]
 
         for user in [BackupUser, MonitorUser, OperatorUser]:
             original_password = self.harness.charm.operator.state.get_user_password(user)
diff --git a/tests/unit/test_mongodb_backups.py b/tests/unit/test_mongodb_backups.py
diff --git a/tests/unit/test_mongodb_provider.py b/tests/unit/test_mongodb_provider.py
diff --git a/tests/unit/test_upgrade.py b/tests/unit/test_upgrade.py
