Merge branch 'devel' into fix_schedules_instance_groups

.github/pr_labeler.yml

@@ -15,5 +15,5 @@
 
 "dependencies":
   - any: ["awx/ui/package.json"]
-  - any: ["awx/requirements/*.txt"]
-  - any: ["awx/requirements/requirements.in"]
+  - any: ["requirements/*.txt"]
+  - any: ["requirements/requirements.in"]

.github/workflows/devel_images.yml

@@ -48,8 +48,11 @@ jobs:
           DEV_DOCKER_TAG_BASE=ghcr.io/${OWNER_LC} COMPOSE_TAG=${GITHUB_REF##*/} make awx-kube-dev-build
           DEV_DOCKER_TAG_BASE=ghcr.io/${OWNER_LC} COMPOSE_TAG=${GITHUB_REF##*/} make awx-kube-build
 
-      - name: Push image
+      - name: Push development images
         run: |
           docker push ghcr.io/${OWNER_LC}/awx_devel:${GITHUB_REF##*/}
           docker push ghcr.io/${OWNER_LC}/awx_kube_devel:${GITHUB_REF##*/}
-          docker push ghcr.io/${OWNER_LC}/awx:${GITHUB_REF##*/}
+
+      - name: Push AWX k8s image, only for upstream and feature branches
+        run: docker push ghcr.io/${OWNER_LC}/awx:${GITHUB_REF##*/}
+        if: endsWith(github.repository, '/awx')

Makefile

@@ -1,6 +1,7 @@
 -include awx/ui_next/Makefile
 
 PYTHON := $(notdir $(shell for i in python3.9 python3; do command -v $$i; done|sed 1q))
+SHELL := bash
 DOCKER_COMPOSE ?= docker-compose
 OFFICIAL ?= no
 NODE ?= node

awx/api/generics.py

@@ -232,7 +232,8 @@ def finalize_response(self, request, response, *args, **kwargs):
 
         response = super(APIView, self).finalize_response(request, response, *args, **kwargs)
         time_started = getattr(self, 'time_started', None)
-        response['X-API-Product-Version'] = get_awx_version()
+        if request.user.is_authenticated:
+            response['X-API-Product-Version'] = get_awx_version()
         response['X-API-Product-Name'] = server_product_name()
 
         response['X-API-Node'] = settings.CLUSTER_HOST_ID

awx/api/urls/urls.py

@@ -30,7 +30,7 @@
     OAuth2TokenList,
     ApplicationOAuth2TokenList,
     OAuth2ApplicationDetail,
-    # HostMetricSummaryMonthlyList, # It will be enabled in future version of the AWX
+    HostMetricSummaryMonthlyList,
 )
 
 from awx.api.views.bulk import (
@@ -123,8 +123,7 @@
     re_path(r'^constructed_inventories/', include(constructed_inventory_urls)),
     re_path(r'^hosts/', include(host_urls)),
     re_path(r'^host_metrics/', include(host_metric_urls)),
-    # It will be enabled in future version of the AWX
-    # re_path(r'^host_metric_summary_monthly/$', HostMetricSummaryMonthlyList.as_view(), name='host_metric_summary_monthly_list'),
+    re_path(r'^host_metric_summary_monthly/$', HostMetricSummaryMonthlyList.as_view(), name='host_metric_summary_monthly_list'),
     re_path(r'^groups/', include(group_urls)),
     re_path(r'^inventory_sources/', include(inventory_source_urls)),
     re_path(r'^inventory_updates/', include(inventory_update_urls)),

awx/api/views/__init__.py

@@ -1564,16 +1564,15 @@ def delete(self, request, *args, **kwargs):
         return Response(status=status.HTTP_204_NO_CONTENT)
 
 
-# It will be enabled in future version of the AWX
-# class HostMetricSummaryMonthlyList(ListAPIView):
-#     name = _("Host Metrics Summary Monthly")
-#     model = models.HostMetricSummaryMonthly
-#     serializer_class = serializers.HostMetricSummaryMonthlySerializer
-#     permission_classes = (IsSystemAdminOrAuditor,)
-#     search_fields = ('date',)
-#
-#     def get_queryset(self):
-#         return self.model.objects.all()
+class HostMetricSummaryMonthlyList(ListAPIView):
+    name = _("Host Metrics Summary Monthly")
+    model = models.HostMetricSummaryMonthly
+    serializer_class = serializers.HostMetricSummaryMonthlySerializer
+    permission_classes = (IsSystemAdminOrAuditor,)
+    search_fields = ('date',)
+
+    def get_queryset(self):
+        return self.model.objects.all()
 
 
 class HostList(HostRelatedSearchMixin, ListCreateAPIView):

awx/api/views/root.py

@@ -107,8 +107,7 @@ def get(self, request, format=None):
         data['groups'] = reverse('api:group_list', request=request)
         data['hosts'] = reverse('api:host_list', request=request)
         data['host_metrics'] = reverse('api:host_metric_list', request=request)
-        # It will be enabled in future version of the AWX
-        # data['host_metric_summary_monthly'] = reverse('api:host_metric_summary_monthly_list', request=request)
+        data['host_metric_summary_monthly'] = reverse('api:host_metric_summary_monthly_list', request=request)
         data['job_templates'] = reverse('api:job_template_list', request=request)
         data['jobs'] = reverse('api:job_list', request=request)
         data['ad_hoc_commands'] = reverse('api:ad_hoc_command_list', request=request)

awx/main/cache.py

@@ -0,0 +1,87 @@
+import functools
+
+from django.conf import settings
+from django.core.cache.backends.base import DEFAULT_TIMEOUT
+from django.core.cache.backends.redis import RedisCache
+
+from redis.exceptions import ConnectionError, ResponseError, TimeoutError
+import socket
+
+# This list comes from what django-redis ignores and the behavior we are trying
+# to retain while dropping the dependency on django-redis.
+IGNORED_EXCEPTIONS = (TimeoutError, ResponseError, ConnectionError, socket.timeout)
+
+CONNECTION_INTERRUPTED_SENTINEL = object()
+
+
+def optionally_ignore_exceptions(func=None, return_value=None):
+    if func is None:
+        return functools.partial(optionally_ignore_exceptions, return_value=return_value)
+
+    @functools.wraps(func)
+    def wrapper(*args, **kwargs):
+        try:
+            return func(*args, **kwargs)
+        except IGNORED_EXCEPTIONS as e:
+            if settings.DJANGO_REDIS_IGNORE_EXCEPTIONS:
+                return return_value
+            raise e.__cause__ or e
+
+    return wrapper
+
+
+class AWXRedisCache(RedisCache):
+    """
+    We just want to wrap the upstream RedisCache class so that we can ignore
+    the exceptions that it raises when the cache is unavailable.
+    """
+
+    @optionally_ignore_exceptions
+    def add(self, key, value, timeout=DEFAULT_TIMEOUT, version=None):
+        return super().add(key, value, timeout, version)
+
+    @optionally_ignore_exceptions(return_value=CONNECTION_INTERRUPTED_SENTINEL)
+    def _get(self, key, default=None, version=None):
+        return super().get(key, default, version)
+
+    def get(self, key, default=None, version=None):
+        value = self._get(key, default, version)
+        if value is CONNECTION_INTERRUPTED_SENTINEL:
+            return default
+        return value
+
+    @optionally_ignore_exceptions
+    def set(self, key, value, timeout=DEFAULT_TIMEOUT, version=None):
+        return super().set(key, value, timeout, version)
+
+    @optionally_ignore_exceptions
+    def touch(self, key, timeout=DEFAULT_TIMEOUT, version=None):
+        return super().touch(key, timeout, version)
+
+    @optionally_ignore_exceptions
+    def delete(self, key, version=None):
+        return super().delete(key, version)
+
+    @optionally_ignore_exceptions
+    def get_many(self, keys, version=None):
+        return super().get_many(keys, version)
+
+    @optionally_ignore_exceptions
+    def has_key(self, key, version=None):
+        return super().has_key(key, version)
+
+    @optionally_ignore_exceptions
+    def incr(self, key, delta=1, version=None):
+        return super().incr(key, delta, version)
+
+    @optionally_ignore_exceptions
+    def set_many(self, data, timeout=DEFAULT_TIMEOUT, version=None):
+        return super().set_many(data, timeout, version)
+
+    @optionally_ignore_exceptions
+    def delete_many(self, keys, version=None):
+        return super().delete_many(keys, version)
+
+    @optionally_ignore_exceptions
+    def clear(self):
+        return super().clear()

awx/main/conf.py

@@ -862,6 +862,15 @@
     category_slug='system',
 )
 
+register(
+    'HOST_METRIC_SUMMARY_TASK_LAST_TS',
+    field_class=fields.DateTimeField,
+    label=_('Last computing date of HostMetricSummaryMonthly'),
+    allow_null=True,
+    category=_('System'),
+    category_slug='system',
+)
+
 register(
     'AWX_CLEANUP_PATHS',
     field_class=fields.BooleanField,

awx/main/dispatch/__init__.py

@@ -55,6 +55,23 @@ def notify(self, channel, payload):
         with self.conn.cursor() as cur:
             cur.execute('SELECT pg_notify(%s, %s);', (channel, payload))
 
+    @staticmethod
+    def current_notifies(conn):
+        """
+        Altered version of .notifies method from psycopg library
+        This removes the outer while True loop so that we only process
+        queued notifications
+        """
+        with conn.lock:
+            try:
+                ns = conn.wait(psycopg.generators.notifies(conn.pgconn))
+            except psycopg.errors._NO_TRACEBACK as ex:
+                raise ex.with_traceback(None)
+        enc = psycopg._encodings.pgconn_encoding(conn.pgconn)
+        for pgn in ns:
+            n = psycopg.connection.Notify(pgn.relname.decode(enc), pgn.extra.decode(enc), pgn.be_pid)
+            yield n
+
     def events(self, select_timeout=5, yield_timeouts=False):
         if not self.conn.autocommit:
             raise RuntimeError('Listening for events can only be done in autocommit mode')
@@ -64,7 +81,7 @@ def events(self, select_timeout=5, yield_timeouts=False):
                 if yield_timeouts:
                     yield None
             else:
-                notification_generator = self.conn.notifies()
+                notification_generator = self.current_notifies(self.conn)
                 for notification in notification_generator:
                     yield notification
 

awx/main/dispatch/pool.py

@@ -417,16 +417,16 @@ def cleanup(self):
                 # the task manager to never do more work
                 current_task = w.current_task
                 if current_task and isinstance(current_task, dict):
-                    endings = ['tasks.task_manager', 'tasks.dependency_manager', 'tasks.workflow_manager']
+                    endings = ('tasks.task_manager', 'tasks.dependency_manager', 'tasks.workflow_manager')
                     current_task_name = current_task.get('task', '')
-                    if any(current_task_name.endswith(e) for e in endings):
+                    if current_task_name.endswith(endings):
                         if 'started' not in current_task:
                             w.managed_tasks[current_task['uuid']]['started'] = time.time()
                         age = time.time() - current_task['started']
                         w.managed_tasks[current_task['uuid']]['age'] = age
                         if age > self.task_manager_timeout:
-                            logger.error(f'{current_task_name} has held the advisory lock for {age}, sending SIGTERM to {w.pid}')
-                            os.kill(w.pid, signal.SIGTERM)
+                            logger.error(f'{current_task_name} has held the advisory lock for {age}, sending SIGUSR1 to {w.pid}')
+                            os.kill(w.pid, signal.SIGUSR1)
 
         for m in orphaned:
             # if all the workers are dead, spawn at least one

awx/main/dispatch/worker/base.py

@@ -121,10 +121,9 @@ def record_statistics(self):
         if time.time() - self.last_stats > 1:  # buffer stat recording to once per second
             try:
                 self.redis.set(f'awx_{self.name}_statistics', self.pool.debug())
-                self.last_stats = time.time()
             except Exception:
                 logger.exception(f"encountered an error communicating with redis to store {self.name} statistics")
-                self.last_stats = time.time()
+            self.last_stats = time.time()
 
     def run(self, *args, **kwargs):
         signal.signal(signal.SIGINT, self.stop)
@@ -175,9 +174,12 @@ def run_periodic_tasks(self):
 
         # record subsystem metrics for the dispatcher
         if current_time - self.last_metrics_gather > 20:
-            self.pool.produce_subsystem_metrics(self.subsystem_metrics)
-            self.subsystem_metrics.set('dispatcher_availability', self.listen_cumulative_time / (current_time - self.last_metrics_gather))
-            self.subsystem_metrics.pipe_execute()
+            try:
+                self.pool.produce_subsystem_metrics(self.subsystem_metrics)
+                self.subsystem_metrics.set('dispatcher_availability', self.listen_cumulative_time / (current_time - self.last_metrics_gather))
+                self.subsystem_metrics.pipe_execute()
+            except Exception:
+                logger.exception(f"encountered an error trying to store {self.name} metrics")
             self.listen_cumulative_time = 0.0
             self.last_metrics_gather = current_time
 
@@ -250,8 +252,8 @@ def work_loop(self, queue, finished, idx, *args):
                     break
             except QueueEmpty:
                 continue
-            except Exception as e:
-                logger.error("Exception on worker {}, restarting: ".format(idx) + str(e))
+            except Exception:
+                logger.exception("Exception on worker {}, reconnecting: ".format(idx))
                 continue
             try:
                 for conn in db.connections.all():

awx/main/management/commands/cleanup_host_metrics.py

@@ -17,6 +17,6 @@ def handle(self, *args, **options):
         months_ago = options.get('months-ago') or None
 
         if not months_ago:
-            months_ago = getattr(settings, 'CLEANUP_HOST_METRICS_THRESHOLD', 12)
+            months_ago = getattr(settings, 'CLEANUP_HOST_METRICS_SOFT_THRESHOLD', 12)
 
         HostMetric.cleanup_task(months_ago)

awx/main/management/commands/host_metric_summary_monthly.py

@@ -0,0 +1,9 @@
+from django.core.management.base import BaseCommand
+from awx.main.tasks.host_metrics import HostMetricSummaryMonthlyTask
+
+
+class Command(BaseCommand):
+    help = 'Computing of HostMetricSummaryMonthly'
+
+    def handle(self, *args, **options):
+        HostMetricSummaryMonthlyTask().execute()

awx/main/models/inventory.py

@@ -899,18 +899,18 @@ def cleanup_task(cls, months_ago):
 
             last_automation_before = now() - dateutil.relativedelta.relativedelta(months=months_ago)
 
-            logger.info(f'Cleanup [HostMetric]: soft-deleting records last automated before {last_automation_before}')
+            logger.info(f'cleanup_host_metrics: soft-deleting records last automated before {last_automation_before}')
             HostMetric.active_objects.filter(last_automation__lt=last_automation_before).update(
                 deleted=True, deleted_counter=models.F('deleted_counter') + 1, last_deleted=now()
             )
             settings.CLEANUP_HOST_METRICS_LAST_TS = now()
         except (TypeError, ValueError):
-            logger.error(f"Cleanup [HostMetric]: months_ago({months_ago}) has to be a positive integer value")
+            logger.error(f"cleanup_host_metrics: months_ago({months_ago}) has to be a positive integer value")
 
 
 class HostMetricSummaryMonthly(models.Model):
     """
-    HostMetric summaries computed by scheduled task <TODO> monthly
+    HostMetric summaries computed by scheduled task 'awx.main.tasks.system.host_metric_summary_monthly' monthly
     """
 
     date = models.DateField(unique=True)