Merge pull request #273 from grafana/dev

Merge dev to main

Author: vstpme

CHANGELOG.md

@@ -1,8 +1,9 @@
 # Change Log
 
-## v1.0.8 (2022-07-21)
+## v1.0.9 (2022-07-21)
 - Frontend bug fixes & improvements
 - Support regex_replace() in templates
+- Bring back alert group caching and list view
 
 ## v1.0.7 (2022-07-18)
 - Backend & frontend bug fixes

engine/apps/alerts/incident_appearance/renderers/base_renderer.py

@@ -18,12 +18,9 @@ def templater_class(self):
 
 
 class AlertGroupBaseRenderer(ABC):
-    def __init__(self, alert_group, alert=None):
-        if alert is None:
-            alert = alert_group.alerts.first()
-
+    def __init__(self, alert_group):
         self.alert_group = alert_group
-        self.alert_renderer = self.alert_renderer_class(alert)
+        self.alert_renderer = self.alert_renderer_class(self.alert_group.alerts.first())
 
     @property
     @abstractmethod

engine/apps/alerts/incident_appearance/renderers/classic_markdown_renderer.py

@@ -20,11 +20,11 @@ def render(self):
 
 
 class AlertGroupClassicMarkdownRenderer(AlertGroupBaseRenderer):
-    def __init__(self, alert_group, alert=None):
-        if alert is None:
-            alert = alert_group.alerts.last()
+    def __init__(self, alert_group):
+        super().__init__(alert_group)
 
-        super().__init__(alert_group, alert)
+        # use the last alert to render content
+        self.alert_renderer = self.alert_renderer_class(self.alert_group.alerts.last())
 
     @property
     def alert_renderer_class(self):

engine/apps/alerts/incident_appearance/renderers/web_renderer.py

@@ -20,11 +20,11 @@ def render(self):
 
 
 class AlertGroupWebRenderer(AlertGroupBaseRenderer):
-    def __init__(self, alert_group, alert=None):
-        if alert is None:
-            alert = alert_group.alerts.last()
+    def __init__(self, alert_group):
+        super().__init__(alert_group)
 
-        super().__init__(alert_group, alert)
+        # use the last alert to render content
+        self.alert_renderer = self.alert_renderer_class(self.alert_group.alerts.last())
 
     @property
     def alert_renderer_class(self):

engine/apps/alerts/models/alert.py

@@ -5,7 +5,7 @@
 from django.apps import apps
 from django.conf import settings
 from django.core.validators import MinLengthValidator
-from django.db import models
+from django.db import models, transaction
 from django.db.models import JSONField
 from django.db.models.signals import post_save
 
@@ -261,6 +261,9 @@ def listen_for_alert_model_save(sender, instance, created, *args, **kwargs):
         else:
             distribute_alert.apply_async((instance.pk,), countdown=TASK_DELAY_SECONDS)
 
+    logger.info(f"Recalculate AG cache. Reason: save alert model {instance.pk}")
+    transaction.on_commit(instance.group.schedule_cache_for_web)
+
 
 # Connect signal to  base Alert class
 post_save.connect(listen_for_alert_model_save, Alert)

engine/apps/alerts/models/alert_group.py

@@ -8,9 +8,12 @@
 from celery import uuid as celery_uuid
 from django.apps import apps
 from django.conf import settings
+from django.core.cache import cache
 from django.core.validators import MinLengthValidator
-from django.db import IntegrityError, models
+from django.db import IntegrityError, models, transaction
 from django.db.models import JSONField, Q, QuerySet
+from django.db.models.signals import post_save
+from django.dispatch import receiver
 from django.utils import timezone
 from django.utils.functional import cached_property
 
@@ -19,9 +22,16 @@
 from apps.alerts.incident_appearance.renderers.slack_renderer import AlertGroupSlackRenderer
 from apps.alerts.incident_log_builder import IncidentLogBuilder
 from apps.alerts.signals import alert_group_action_triggered_signal
-from apps.alerts.tasks import acknowledge_reminder_task, call_ack_url, send_alert_group_signal, unsilence_task
+from apps.alerts.tasks import (
+    acknowledge_reminder_task,
+    call_ack_url,
+    schedule_cache_for_alert_group,
+    send_alert_group_signal,
+    unsilence_task,
+)
 from apps.slack.slack_formatter import SlackFormatter
 from apps.user_management.models import User
+from common.mixins.use_random_readonly_db_manager_mixin import UseRandomReadonlyDbManagerMixin
 from common.public_primary_keys import generate_public_primary_key, increase_public_primary_key_length
 from common.utils import clean_markup, str_or_backup
 
@@ -98,6 +108,10 @@ def filter(self, *args, **kwargs):
         return super().filter(*args, **kwargs, is_archived=False)
 
 
+class AlertGroupManager(UseRandomReadonlyDbManagerMixin, models.Manager):
+    pass
+
+
 class AlertGroupSlackRenderingMixin:
     """
     Ideally this mixin should not exist. Instead of this instance of AlertGroupSlackRenderer should be created and used
@@ -120,8 +134,8 @@ def slack_templated_first_alert(self):
 
 
 class AlertGroup(AlertGroupSlackRenderingMixin, EscalationSnapshotMixin, models.Model):
-    all_objects = AlertGroupQuerySet.as_manager()
-    unarchived_objects = UnarchivedAlertGroupQuerySet.as_manager()
+    all_objects = AlertGroupManager.from_queryset(AlertGroupQuerySet)()
+    unarchived_objects = AlertGroupManager.from_queryset(UnarchivedAlertGroupQuerySet)()
 
     (
         NEW,
@@ -228,6 +242,8 @@ class AlertGroup(AlertGroupSlackRenderingMixin, EscalationSnapshotMixin, models.
 
     active_escalation_id = models.CharField(max_length=100, null=True, default=None)  # ID generated by celery
     active_resolve_calculation_id = models.CharField(max_length=100, null=True, default=None)  # ID generated by celery
+    # ID generated by celery
+    active_cache_for_web_calculation_id = models.CharField(max_length=100, null=True, default=None)
 
     SILENCE_DELAY_OPTIONS = (
         (1800, "30 minutes"),
@@ -299,9 +315,7 @@ def status(self):
         related_name="dependent_alert_groups",
     )
 
-    # cached_render_for_web and active_cache_for_web_calculation_id are deprecated
-    cached_render_for_web = models.JSONField(default=dict)
-    active_cache_for_web_calculation_id = models.CharField(max_length=100, null=True, default=None)
+    cached_render_for_web = JSONField(default=dict)
 
     last_unique_unacknowledge_process_id = models.CharField(max_length=100, null=True, default=None)
     is_archived = models.BooleanField(default=False)
@@ -390,6 +404,18 @@ def skip_escalation_in_slack(self):
     def is_alert_a_resolve_signal(self, alert):
         raise NotImplementedError
 
+    def cache_for_web(self, organization):
+        from apps.api.serializers.alert_group import AlertGroupSerializer
+
+        # Re-take object to switch connection from readonly db to master.
+        _self = AlertGroup.all_objects.get(pk=self.pk)
+        _self.cached_render_for_web = AlertGroupSerializer(self, context={"organization": organization}).data
+        self.cached_render_for_web = _self.cached_render_for_web
+        _self.save(update_fields=["cached_render_for_web"])
+
+    def schedule_cache_for_web(self):
+        schedule_cache_for_alert_group.apply_async((self.pk,))
+
     @property
     def permalink(self):
         if self.slack_message is not None:
@@ -399,6 +425,10 @@ def permalink(self):
     def web_link(self):
         return urljoin(self.channel.organization.web_link, f"?page=incident&id={self.public_primary_key}")
 
+    @property
+    def alerts_count(self):
+        return self.alerts.count()
+
     @property
     def happened_while_maintenance(self):
         return self.root_alert_group is not None and self.root_alert_group.maintenance_uuid is not None
@@ -419,6 +449,10 @@ def acknowledge_by_user(self, user: User, action_source: Optional[str] = None) -
             self.unresolve()
             self.log_records.create(type=AlertGroupLogRecord.TYPE_UN_RESOLVED, author=user, reason="Acknowledge button")
 
+            # clear resolve report cache
+            cache_key = "render_after_resolve_report_json_{}".format(self.pk)
+            cache.delete(cache_key)
+
         self.acknowledge(acknowledged_by_user=user, acknowledged_by=AlertGroup.USER)
         self.stop_escalation()
         if self.is_root_alert_group:
@@ -639,6 +673,9 @@ def un_resolve_by_user(self, user: User, action_source: Optional[str] = None) ->
             self.unresolve()
             log_record = self.log_records.create(type=AlertGroupLogRecord.TYPE_UN_RESOLVED, author=user)
 
+            # clear resolve report cache
+            self.drop_cached_after_resolve_report_json()
+
             if self.is_root_alert_group:
                 self.start_escalation_if_needed()
 
@@ -811,6 +848,10 @@ def silence_by_user(self, user: User, silence_delay: Optional[int], action_sourc
             self.unresolve()
             self.log_records.create(type=AlertGroupLogRecord.TYPE_UN_RESOLVED, author=user, reason="Silence button")
 
+            # clear resolve report cache
+            cache_key = "render_after_resolve_report_json_{}".format(self.pk)
+            cache.delete(cache_key)
+
         if self.acknowledged:
             self.unacknowledge()
             self.log_records.create(type=AlertGroupLogRecord.TYPE_UN_ACK, author=user, reason="Silence button")
@@ -1019,6 +1060,8 @@ def bulk_acknowledge(user: User, alert_groups: "QuerySet[AlertGroup]") -> None:
                 author=user,
                 reason="Bulk action acknowledge",
             )
+            # clear resolve report cache
+            alert_group.drop_cached_after_resolve_report_json()
 
         for alert_group in alert_groups_to_unsilence_before_acknowledge_list:
             alert_group.log_records.create(
@@ -1151,6 +1194,8 @@ def bulk_restart(user: User, alert_groups: "QuerySet[AlertGroup]") -> None:
                 reason="Bulk action restart",
             )
 
+            alert_group.drop_cached_after_resolve_report_json()
+
             if alert_group.is_root_alert_group:
                 alert_group.start_escalation_if_needed()
 
@@ -1248,6 +1293,7 @@ def bulk_silence(user: User, alert_groups: "QuerySet[AlertGroup]", silence_delay
                 author=user,
                 reason="Bulk action silence",
             )
+            alert_group.drop_cached_after_resolve_report_json()
 
         for alert_group in alert_groups_to_unsilence_before_silence_list:
             alert_group.log_records.create(
@@ -1437,7 +1483,7 @@ def get_acknowledge_text(self, mention_user=False):
         else:
             return "Acknowledged"
 
-    def render_after_resolve_report_json(self):
+    def non_cached_after_resolve_report_json(self):
         AlertGroupLogRecord = apps.get_model("alerts", "AlertGroupLogRecord")
         UserNotificationPolicyLogRecord = apps.get_model("base", "UserNotificationPolicyLogRecord")
         ResolutionNote = apps.get_model("alerts", "ResolutionNote")
@@ -1455,6 +1501,21 @@ def render_after_resolve_report_json(self):
                 result_log_report.append(log_record.render_log_line_json())
         return result_log_report
 
+    def render_after_resolve_report_json(self):
+        cache_key = "render_after_resolve_report_json_{}".format(self.pk)
+
+        # cache.get_or_set in some cases returns None, so use get and set cache methods separately
+        log_report = cache.get(cache_key)
+        if log_report is None:
+            log_report = self.non_cached_after_resolve_report_json()
+            cache.set(cache_key, log_report)
+        return log_report
+
+    def drop_cached_after_resolve_report_json(self):
+        cache_key = "render_after_resolve_report_json_{}".format(self.pk)
+        if cache_key in cache:
+            cache.delete(cache_key)
+
     @property
     def has_resolution_notes(self):
         return self.resolution_notes.exists()
@@ -1534,3 +1595,14 @@ def last_stop_escalation_log(self):
         )
 
         return stop_escalation_log
+
+
+@receiver(post_save, sender=AlertGroup)
+def listen_for_alert_group_model_save(sender, instance, created, *args, **kwargs):
+    if (
+        kwargs is not None
+        and "update_fields" in kwargs
+        and kwargs["update_fields"] is dict
+        and "cached_render_for_web" not in kwargs["update_fields"]
+    ):
+        transaction.on_commit(instance.schedule_cache_for_alert_group)

engine/apps/alerts/models/alert_group_log_record.py

@@ -3,7 +3,7 @@
 
 import humanize
 from django.apps import apps
-from django.db import models
+from django.db import models, transaction
 from django.db.models import JSONField
 from django.db.models.signals import post_save
 from django.dispatch import receiver
@@ -546,6 +546,7 @@ def get_step_specific_info(self):
 
 @receiver(post_save, sender=AlertGroupLogRecord)
 def listen_for_alertgrouplogrecord(sender, instance, created, *args, **kwargs):
+    instance.alert_group.drop_cached_after_resolve_report_json()
     if instance.type != AlertGroupLogRecord.TYPE_DELETED:
         if not instance.alert_group.is_maintenance_incident:
             alert_group_pk = instance.alert_group.pk
@@ -554,3 +555,6 @@ def listen_for_alertgrouplogrecord(sender, instance, created, *args, **kwargs):
                 f"alert group event: {instance.get_type_display()}"
             )
             send_update_log_report_signal.apply_async(kwargs={"alert_group_pk": alert_group_pk}, countdown=8)
+
+    logger.info(f"Recalculate AG cache. Reason: save alert_group_log_record model {instance.pk}")
+    transaction.on_commit(instance.alert_group.schedule_cache_for_web)

engine/apps/alerts/models/alert_receive_channel.py

@@ -19,7 +19,11 @@
 from apps.alerts.grafana_alerting_sync_manager.grafana_alerting_sync import GrafanaAlertingSyncManager
 from apps.alerts.integration_options_mixin import IntegrationOptionsMixin
 from apps.alerts.models.maintainable_object import MaintainableObject
-from apps.alerts.tasks import disable_maintenance, sync_grafana_alerting_contact_points
+from apps.alerts.tasks import (
+    disable_maintenance,
+    invalidate_web_cache_for_alert_group,
+    sync_grafana_alerting_contact_points,
+)
 from apps.base.messaging import get_messaging_backend_from_id
 from apps.base.utils import live_settings
 from apps.integrations.metadata import heartbeat
@@ -689,6 +693,16 @@ def listen_for_alertreceivechannel_model_save(sender, instance, created, *args,
             create_organization_log(
                 instance.organization, None, OrganizationLogType.TYPE_HEARTBEAT_CREATED, description
             )
+    else:
+        logger.info(f"Drop AG cache. Reason: save alert_receive_channel {instance.pk}")
+        if kwargs is not None:
+            if "update_fields" in kwargs:
+                if kwargs["update_fields"] is not None:
+                    # Hack to not to invalidate web cache on AlertReceiveChannel.start_send_rate_limit_message_task
+                    if "rate_limit_message_task_id" in kwargs["update_fields"]:
+                        return
+
+        invalidate_web_cache_for_alert_group.apply_async(kwargs={"channel_pk": instance.pk})
 
     if instance.integration == AlertReceiveChannel.INTEGRATION_GRAFANA_ALERTING:
         if created:

engine/apps/alerts/tasks/__init__.py

@@ -9,7 +9,7 @@
 from .delete_alert_group import delete_alert_group  # noqa: F401
 from .distribute_alert import distribute_alert  # noqa: F401
 from .escalate_alert_group import escalate_alert_group  # noqa: F401
-from .invalidate_web_cache_for_alert_group import invalidate_web_cache_for_alert_group  # noqa: F401, todo: remove
+from .invalidate_web_cache_for_alert_group import invalidate_web_cache_for_alert_group  # noqa: F401
 from .invite_user_to_join_incident import invite_user_to_join_incident  # noqa: F401
 from .maintenance import disable_maintenance  # noqa: F401
 from .notify_all import notify_all_task  # noqa: F401