oncall-engine/engine/apps/alerts/models/alert.py

322 lines
13 KiB
Python
Raw Permalink Normal View History

import hashlib
import logging
import typing
from functools import partial
from uuid import uuid4
from django.conf import settings
from django.core.validators import MinLengthValidator
from django.db import models, transaction
from django.db.models import JSONField
from apps.alerts import tasks
from apps.alerts.constants import TASK_DELAY_SECONDS
from apps.alerts.incident_appearance.templaters import TemplateLoader
from apps.alerts.signals import alert_group_escalation_snapshot_built
from apps.alerts.tasks.distribute_alert import send_alert_create_signal
from apps.labels.alert_group_labels import gather_alert_labels, save_alert_group_labels
Support prescribed labels (#3848) # What this PR does **Cleanup label typing:** 1. LabelParam -> two separate types LabekKey and LabelValue 2. LabelData -> renamed to LabelPair. 3. LabelKeyData -> renamed to LabelOption Data is not giving any info about what this type represents. 4. Remove LabelsData and LabelsKeysData types. They are just list of types listed above and with new naming it feels obsolete. 5. ValueData removed. LabelPair is used instead. 6. Rework AlertGroupCustomLabel to use LabelKey type for key to make type system more consistent. Name model type AlertGroupCustomLabel**DB** and api type AlertGroupCustomLabel**API** to clearly distinguish them. **Split update_labels_cache into two tasks** update_label_option_cache and update_label_pairs_cache. Original task was expecting array of LabelsData (now it's LabelPair) OR one LabelKeyData ( now it's LabelOption). I believe having one function with two sp different argument types makes it more complicated for understanding. **Make OnCall backend support prescribed labels**. OnCall will sync and store "prescribed" field for key and values, so Label dropdown able to disable editing for certain labels. ## Which issue(s) this PR fixes ## Checklist - [x] Unit, integration, and e2e (if applicable) tests updated - [ ] Documentation added (or `pr:no public docs` PR label added if not required) - [x] `CHANGELOG.md` updated (or `pr:no changelog` PR label added if not required) --------- Co-authored-by: Maxim Mordasov <maxim.mordasov@grafana.com> Co-authored-by: Yulya Artyukhina <Ferril.darkdiver@gmail.com>
2024-02-20 14:42:51 +08:00
from apps.labels.types import AlertLabels
Support alert routing based on labels (#3778) # What this PR does This PR adds support for routing alerts based on labels. https://www.loom.com/share/4401de6e3c4945d5b8961fe43ee373c9 Additionally: - improve the typing around the `get_object` method that is inherited by [`PublicPrimaryKeyMixin.get_object`](https://github.com/grafana/oncall/blob/dev/engine/common/api_helpers/mixins.py#L153) in most of our models. `PublicPrimaryKeyMixin` is generic, so it can be more strongly typed when it is being subclassed, which results in better typing of the `get_object` method in child classes - I decided to do this because I started looking into this task via the [`AlertReceiveChannelView.send_demo_alert` method/endpoint](https://github.com/grafana/oncall/blob/dev/engine/apps/api/views/alert_receive_channel.py#L242). Within that method, `instance` is not typed because the inherited `get_object` method is not typed.. I digress 😄 - improve typing around `Alert.create` and `apps.integrations.tasks.create_alert` functions - make `Alert.render_group_data` more DRY by extracting some logic out into `Alert._apply_jinja_template_to_alert_payload_and_labels` - deduplicate the logic of `value.strip().lower() in ["1", "true", "ok"]` into a shared function, `common.jinja_templater.apply_jinja_template.templated_value_is_truthy` Closes https://github.com/grafana/oncall-private/issues/2490 ## Checklist - [x] Unit, integration, and e2e (if applicable) tests updated - [x] `CHANGELOG.md` updated (or `pr:no changelog` PR label added if not required) - [x] Documentation added (or `pr:no public docs` PR label added if not required) (will be done in #3762)
2024-01-30 13:07:19 -05:00
from common.jinja_templater import apply_jinja_template_to_alert_payload_and_labels
from common.jinja_templater.apply_jinja_template import (
JinjaTemplateError,
JinjaTemplateWarning,
templated_value_is_truthy,
)
from common.public_primary_keys import generate_public_primary_key, increase_public_primary_key_length
if typing.TYPE_CHECKING:
2024-10-09 17:46:38 -04:00
from django.db.models.manager import RelatedManager
Support alert routing based on labels (#3778) # What this PR does This PR adds support for routing alerts based on labels. https://www.loom.com/share/4401de6e3c4945d5b8961fe43ee373c9 Additionally: - improve the typing around the `get_object` method that is inherited by [`PublicPrimaryKeyMixin.get_object`](https://github.com/grafana/oncall/blob/dev/engine/common/api_helpers/mixins.py#L153) in most of our models. `PublicPrimaryKeyMixin` is generic, so it can be more strongly typed when it is being subclassed, which results in better typing of the `get_object` method in child classes - I decided to do this because I started looking into this task via the [`AlertReceiveChannelView.send_demo_alert` method/endpoint](https://github.com/grafana/oncall/blob/dev/engine/apps/api/views/alert_receive_channel.py#L242). Within that method, `instance` is not typed because the inherited `get_object` method is not typed.. I digress 😄 - improve typing around `Alert.create` and `apps.integrations.tasks.create_alert` functions - make `Alert.render_group_data` more DRY by extracting some logic out into `Alert._apply_jinja_template_to_alert_payload_and_labels` - deduplicate the logic of `value.strip().lower() in ["1", "true", "ok"]` into a shared function, `common.jinja_templater.apply_jinja_template.templated_value_is_truthy` Closes https://github.com/grafana/oncall-private/issues/2490 ## Checklist - [x] Unit, integration, and e2e (if applicable) tests updated - [x] `CHANGELOG.md` updated (or `pr:no changelog` PR label added if not required) - [x] Documentation added (or `pr:no public docs` PR label added if not required) (will be done in #3762)
2024-01-30 13:07:19 -05:00
from apps.alerts.models import AlertGroup, AlertReceiveChannel, ChannelFilter
logger = logging.getLogger(__name__)
logger.setLevel(logging.DEBUG)
def generate_public_primary_key_for_alert():
prefix = "A"
new_public_primary_key = generate_public_primary_key(prefix)
failure_counter = 0
while Alert.objects.filter(public_primary_key=new_public_primary_key).exists():
new_public_primary_key = increase_public_primary_key_length(
failure_counter=failure_counter, prefix=prefix, model_name="Alert"
)
failure_counter += 1
return new_public_primary_key
class Alert(models.Model):
group: typing.Optional["AlertGroup"]
2024-10-09 17:46:38 -04:00
resolved_alert_groups: "RelatedManager['AlertGroup']"
public_primary_key = models.CharField(
max_length=20,
validators=[MinLengthValidator(settings.PUBLIC_PRIMARY_KEY_MIN_LENGTH + 1)],
unique=True,
default=generate_public_primary_key_for_alert,
)
is_resolve_signal = models.BooleanField(default=False)
is_the_first_alert_in_group = models.BooleanField(default=False)
message = models.TextField(max_length=3000, default=None, null=True)
image_url = models.URLField(default=None, null=True, max_length=300)
delivered = models.BooleanField(default=False)
title = models.TextField(max_length=1500, default=None, null=True)
created_at = models.DateTimeField(auto_now_add=True)
link_to_upstream_details = models.URLField(max_length=500, default=None, null=True)
integration_unique_data = JSONField(default=None, null=True)
raw_request_data = JSONField()
# This hash is for integration-specific needs
integration_optimization_hash = models.CharField(max_length=100, db_index=True, default=None, null=True)
group = models.ForeignKey(
"alerts.AlertGroup", on_delete=models.CASCADE, null=True, default=None, related_name="alerts"
)
Support alert routing based on labels (#3778) # What this PR does This PR adds support for routing alerts based on labels. https://www.loom.com/share/4401de6e3c4945d5b8961fe43ee373c9 Additionally: - improve the typing around the `get_object` method that is inherited by [`PublicPrimaryKeyMixin.get_object`](https://github.com/grafana/oncall/blob/dev/engine/common/api_helpers/mixins.py#L153) in most of our models. `PublicPrimaryKeyMixin` is generic, so it can be more strongly typed when it is being subclassed, which results in better typing of the `get_object` method in child classes - I decided to do this because I started looking into this task via the [`AlertReceiveChannelView.send_demo_alert` method/endpoint](https://github.com/grafana/oncall/blob/dev/engine/apps/api/views/alert_receive_channel.py#L242). Within that method, `instance` is not typed because the inherited `get_object` method is not typed.. I digress 😄 - improve typing around `Alert.create` and `apps.integrations.tasks.create_alert` functions - make `Alert.render_group_data` more DRY by extracting some logic out into `Alert._apply_jinja_template_to_alert_payload_and_labels` - deduplicate the logic of `value.strip().lower() in ["1", "true", "ok"]` into a shared function, `common.jinja_templater.apply_jinja_template.templated_value_is_truthy` Closes https://github.com/grafana/oncall-private/issues/2490 ## Checklist - [x] Unit, integration, and e2e (if applicable) tests updated - [x] `CHANGELOG.md` updated (or `pr:no changelog` PR label added if not required) - [x] Documentation added (or `pr:no public docs` PR label added if not required) (will be done in #3762)
2024-01-30 13:07:19 -05:00
RawRequestData: typing.TypeAlias = typing.Union[typing.Dict, typing.List]
def get_integration_optimization_hash(self):
"""
Should be overloaded in child classes.
"""
raise NotImplementedError
@classmethod
def create(
cls,
Support alert routing based on labels (#3778) # What this PR does This PR adds support for routing alerts based on labels. https://www.loom.com/share/4401de6e3c4945d5b8961fe43ee373c9 Additionally: - improve the typing around the `get_object` method that is inherited by [`PublicPrimaryKeyMixin.get_object`](https://github.com/grafana/oncall/blob/dev/engine/common/api_helpers/mixins.py#L153) in most of our models. `PublicPrimaryKeyMixin` is generic, so it can be more strongly typed when it is being subclassed, which results in better typing of the `get_object` method in child classes - I decided to do this because I started looking into this task via the [`AlertReceiveChannelView.send_demo_alert` method/endpoint](https://github.com/grafana/oncall/blob/dev/engine/apps/api/views/alert_receive_channel.py#L242). Within that method, `instance` is not typed because the inherited `get_object` method is not typed.. I digress 😄 - improve typing around `Alert.create` and `apps.integrations.tasks.create_alert` functions - make `Alert.render_group_data` more DRY by extracting some logic out into `Alert._apply_jinja_template_to_alert_payload_and_labels` - deduplicate the logic of `value.strip().lower() in ["1", "true", "ok"]` into a shared function, `common.jinja_templater.apply_jinja_template.templated_value_is_truthy` Closes https://github.com/grafana/oncall-private/issues/2490 ## Checklist - [x] Unit, integration, and e2e (if applicable) tests updated - [x] `CHANGELOG.md` updated (or `pr:no changelog` PR label added if not required) - [x] Documentation added (or `pr:no public docs` PR label added if not required) (will be done in #3762)
2024-01-30 13:07:19 -05:00
title: typing.Optional[str],
message: typing.Optional[str],
image_url: typing.Optional[str],
link_to_upstream_details: typing.Optional[str],
alert_receive_channel: "AlertReceiveChannel",
integration_unique_data: typing.Optional[typing.Dict],
raw_request_data: RawRequestData,
enable_autoresolve=True,
Support alert routing based on labels (#3778) # What this PR does This PR adds support for routing alerts based on labels. https://www.loom.com/share/4401de6e3c4945d5b8961fe43ee373c9 Additionally: - improve the typing around the `get_object` method that is inherited by [`PublicPrimaryKeyMixin.get_object`](https://github.com/grafana/oncall/blob/dev/engine/common/api_helpers/mixins.py#L153) in most of our models. `PublicPrimaryKeyMixin` is generic, so it can be more strongly typed when it is being subclassed, which results in better typing of the `get_object` method in child classes - I decided to do this because I started looking into this task via the [`AlertReceiveChannelView.send_demo_alert` method/endpoint](https://github.com/grafana/oncall/blob/dev/engine/apps/api/views/alert_receive_channel.py#L242). Within that method, `instance` is not typed because the inherited `get_object` method is not typed.. I digress 😄 - improve typing around `Alert.create` and `apps.integrations.tasks.create_alert` functions - make `Alert.render_group_data` more DRY by extracting some logic out into `Alert._apply_jinja_template_to_alert_payload_and_labels` - deduplicate the logic of `value.strip().lower() in ["1", "true", "ok"]` into a shared function, `common.jinja_templater.apply_jinja_template.templated_value_is_truthy` Closes https://github.com/grafana/oncall-private/issues/2490 ## Checklist - [x] Unit, integration, and e2e (if applicable) tests updated - [x] `CHANGELOG.md` updated (or `pr:no changelog` PR label added if not required) - [x] Documentation added (or `pr:no public docs` PR label added if not required) (will be done in #3762)
2024-01-30 13:07:19 -05:00
is_demo: bool = False,
channel_filter: typing.Optional["ChannelFilter"] = None,
received_at: typing.Optional[str] = None,
) -> "Alert":
"""
Creates an alert and a group if needed.
"""
# This import is here to avoid circular imports
`apps.get_model` -> `import` (#2619) # What this PR does Remove [`apps.get_model`](https://docs.djangoproject.com/en/3.2/ref/applications/#django.apps.apps.get_model) invocations and use inline `import` statements in places where models are imported within functions/methods to avoid circular imports. I believe `import` statements are more appropriate for most use cases as they allow for better static code analysis & formatting, and solve the issue of circular imports without being unnecessarily dynamic as `apps.get_model`. With `import` statements, it's possible to: - Jump to model definitions in most IDEs - Automatically sort inline imports with `isort` - Find import errors faster/easier (most IDEs highlight broken imports) - Have more consistency across regular & inline imports when importing models This PR also adds a flake8 rule to ban imports of `django.apps.apps`, so it's harder to use `apps.get_model` by mistake (it's possible to ignore this rule by using `# noqa: I251`). The rule is not enforced on directories with migration files, because `apps.get_model` is often used to get a historical state of a model, which is useful when writing migrations ([see this SO answer for more details](https://stackoverflow.com/a/37769213)). So `apps.get_model` is considered OK in migrations (even necessary in some cases). ## Checklist - [x] Unit, integration, and e2e (if applicable) tests updated - [x] Documentation added (or `pr:no public docs` PR label added if not required) - [x] `CHANGELOG.md` updated (or `pr:no changelog` PR label added if not required)
2023-07-25 10:43:23 +01:00
from apps.alerts.models import AlertGroup, AlertGroupLogRecord, AlertReceiveChannel, ChannelFilter
alert_labels = gather_alert_labels(alert_receive_channel, raw_request_data)
group_data = Alert.render_group_data(alert_receive_channel, raw_request_data, alert_labels, is_demo)
Support alert routing based on labels (#3778) # What this PR does This PR adds support for routing alerts based on labels. https://www.loom.com/share/4401de6e3c4945d5b8961fe43ee373c9 Additionally: - improve the typing around the `get_object` method that is inherited by [`PublicPrimaryKeyMixin.get_object`](https://github.com/grafana/oncall/blob/dev/engine/common/api_helpers/mixins.py#L153) in most of our models. `PublicPrimaryKeyMixin` is generic, so it can be more strongly typed when it is being subclassed, which results in better typing of the `get_object` method in child classes - I decided to do this because I started looking into this task via the [`AlertReceiveChannelView.send_demo_alert` method/endpoint](https://github.com/grafana/oncall/blob/dev/engine/apps/api/views/alert_receive_channel.py#L242). Within that method, `instance` is not typed because the inherited `get_object` method is not typed.. I digress 😄 - improve typing around `Alert.create` and `apps.integrations.tasks.create_alert` functions - make `Alert.render_group_data` more DRY by extracting some logic out into `Alert._apply_jinja_template_to_alert_payload_and_labels` - deduplicate the logic of `value.strip().lower() in ["1", "true", "ok"]` into a shared function, `common.jinja_templater.apply_jinja_template.templated_value_is_truthy` Closes https://github.com/grafana/oncall-private/issues/2490 ## Checklist - [x] Unit, integration, and e2e (if applicable) tests updated - [x] `CHANGELOG.md` updated (or `pr:no changelog` PR label added if not required) - [x] Documentation added (or `pr:no public docs` PR label added if not required) (will be done in #3762)
2024-01-30 13:07:19 -05:00
if channel_filter is None:
channel_filter = ChannelFilter.select_filter(alert_receive_channel, raw_request_data, alert_labels)
# Get or create group
group, group_created = AlertGroup.objects.get_or_create_grouping(
channel=alert_receive_channel,
channel_filter=channel_filter,
group_data=group_data,
received_at=received_at,
)
logger.debug(f"alert group {group.pk} created={group_created}")
# Create alert
alert = cls(
is_resolve_signal=group_data.is_resolve_signal,
title=title,
message=message,
image_url=image_url,
link_to_upstream_details=link_to_upstream_details,
group=group,
integration_unique_data=integration_unique_data,
raw_request_data=raw_request_data,
is_the_first_alert_in_group=group_created,
)
alert.save()
2024-04-16 08:39:00 -06:00
logger.debug(f"alert {alert.pk} created for alert group {group.pk}")
transaction.on_commit(partial(send_alert_create_signal.apply_async, (alert.pk,)))
if group_created:
save_alert_group_labels(group, alert_receive_channel, alert_labels)
group.log_records.create(type=AlertGroupLogRecord.TYPE_REGISTERED)
group.log_records.create(type=AlertGroupLogRecord.TYPE_ROUTE_ASSIGNED)
if group_created or alert.group.pause_escalation:
# Build escalation snapshot if needed and start escalation
alert.group.start_escalation_if_needed(countdown=TASK_DELAY_SECONDS)
if group_created:
# TODO: consider moving to start_escalation_if_needed
alert_group_escalation_snapshot_built.send(sender=cls.__class__, alert_group=alert.group)
mark_as_acknowledged = group_data.is_acknowledge_signal
if not group.acknowledged and mark_as_acknowledged:
group.acknowledge_by_source()
mark_as_resolved = (
enable_autoresolve and group_data.is_resolve_signal and alert_receive_channel.allow_source_based_resolving
)
if not group.resolved and mark_as_resolved:
group.resolve_by_source()
if group_created:
# all code below related to maintenance mode
maintenance_uuid = None
if alert_receive_channel.maintenance_mode == AlertReceiveChannel.MAINTENANCE:
maintenance_uuid = alert_receive_channel.maintenance_uuid
if maintenance_uuid is not None:
try:
maintenance_incident = AlertGroup.objects.get(maintenance_uuid=maintenance_uuid)
group.root_alert_group = maintenance_incident
group.save(update_fields=["root_alert_group"])
log_record_for_root_incident = maintenance_incident.log_records.create(
type=AlertGroupLogRecord.TYPE_ATTACHED, dependent_alert_group=group, reason="Attach dropdown"
)
logger.debug(
f"call send_alert_group_signal for alert_group {maintenance_incident.pk} (maintenance), "
f"log record {log_record_for_root_incident.pk} with type "
f"'{log_record_for_root_incident.get_type_display()}'"
)
transaction.on_commit(partial(tasks.send_alert_group_signal.delay, log_record_for_root_incident.pk))
except AlertGroup.DoesNotExist:
pass
return alert
def wipe(self, wiped_by, wiped_at):
wiped_by_user_verbal = "by " + wiped_by.username
self.integration_unique_data = {}
self.raw_request_data = {}
self.title = f"Wiped {wiped_by_user_verbal} at {wiped_at.strftime('%Y-%m-%d')}"
self.message = ""
self.image_url = None
self.link_to_upstream_details = None
self.save(
update_fields=[
"integration_unique_data",
"raw_request_data",
"title",
"message",
"image_url",
"link_to_upstream_details",
]
)
@classmethod
Support alert routing based on labels (#3778) # What this PR does This PR adds support for routing alerts based on labels. https://www.loom.com/share/4401de6e3c4945d5b8961fe43ee373c9 Additionally: - improve the typing around the `get_object` method that is inherited by [`PublicPrimaryKeyMixin.get_object`](https://github.com/grafana/oncall/blob/dev/engine/common/api_helpers/mixins.py#L153) in most of our models. `PublicPrimaryKeyMixin` is generic, so it can be more strongly typed when it is being subclassed, which results in better typing of the `get_object` method in child classes - I decided to do this because I started looking into this task via the [`AlertReceiveChannelView.send_demo_alert` method/endpoint](https://github.com/grafana/oncall/blob/dev/engine/apps/api/views/alert_receive_channel.py#L242). Within that method, `instance` is not typed because the inherited `get_object` method is not typed.. I digress 😄 - improve typing around `Alert.create` and `apps.integrations.tasks.create_alert` functions - make `Alert.render_group_data` more DRY by extracting some logic out into `Alert._apply_jinja_template_to_alert_payload_and_labels` - deduplicate the logic of `value.strip().lower() in ["1", "true", "ok"]` into a shared function, `common.jinja_templater.apply_jinja_template.templated_value_is_truthy` Closes https://github.com/grafana/oncall-private/issues/2490 ## Checklist - [x] Unit, integration, and e2e (if applicable) tests updated - [x] `CHANGELOG.md` updated (or `pr:no changelog` PR label added if not required) - [x] Documentation added (or `pr:no public docs` PR label added if not required) (will be done in #3762)
2024-01-30 13:07:19 -05:00
def _apply_jinja_template_to_alert_payload_and_labels(
cls,
template: str,
template_name: str,
alert_receive_channel: "AlertReceiveChannel",
raw_request_data: RawRequestData,
Support prescribed labels (#3848) # What this PR does **Cleanup label typing:** 1. LabelParam -> two separate types LabekKey and LabelValue 2. LabelData -> renamed to LabelPair. 3. LabelKeyData -> renamed to LabelOption Data is not giving any info about what this type represents. 4. Remove LabelsData and LabelsKeysData types. They are just list of types listed above and with new naming it feels obsolete. 5. ValueData removed. LabelPair is used instead. 6. Rework AlertGroupCustomLabel to use LabelKey type for key to make type system more consistent. Name model type AlertGroupCustomLabel**DB** and api type AlertGroupCustomLabel**API** to clearly distinguish them. **Split update_labels_cache into two tasks** update_label_option_cache and update_label_pairs_cache. Original task was expecting array of LabelsData (now it's LabelPair) OR one LabelKeyData ( now it's LabelOption). I believe having one function with two sp different argument types makes it more complicated for understanding. **Make OnCall backend support prescribed labels**. OnCall will sync and store "prescribed" field for key and values, so Label dropdown able to disable editing for certain labels. ## Which issue(s) this PR fixes ## Checklist - [x] Unit, integration, and e2e (if applicable) tests updated - [ ] Documentation added (or `pr:no public docs` PR label added if not required) - [x] `CHANGELOG.md` updated (or `pr:no changelog` PR label added if not required) --------- Co-authored-by: Maxim Mordasov <maxim.mordasov@grafana.com> Co-authored-by: Yulya Artyukhina <Ferril.darkdiver@gmail.com>
2024-02-20 14:42:51 +08:00
labels: typing.Optional[AlertLabels],
Support alert routing based on labels (#3778) # What this PR does This PR adds support for routing alerts based on labels. https://www.loom.com/share/4401de6e3c4945d5b8961fe43ee373c9 Additionally: - improve the typing around the `get_object` method that is inherited by [`PublicPrimaryKeyMixin.get_object`](https://github.com/grafana/oncall/blob/dev/engine/common/api_helpers/mixins.py#L153) in most of our models. `PublicPrimaryKeyMixin` is generic, so it can be more strongly typed when it is being subclassed, which results in better typing of the `get_object` method in child classes - I decided to do this because I started looking into this task via the [`AlertReceiveChannelView.send_demo_alert` method/endpoint](https://github.com/grafana/oncall/blob/dev/engine/apps/api/views/alert_receive_channel.py#L242). Within that method, `instance` is not typed because the inherited `get_object` method is not typed.. I digress 😄 - improve typing around `Alert.create` and `apps.integrations.tasks.create_alert` functions - make `Alert.render_group_data` more DRY by extracting some logic out into `Alert._apply_jinja_template_to_alert_payload_and_labels` - deduplicate the logic of `value.strip().lower() in ["1", "true", "ok"]` into a shared function, `common.jinja_templater.apply_jinja_template.templated_value_is_truthy` Closes https://github.com/grafana/oncall-private/issues/2490 ## Checklist - [x] Unit, integration, and e2e (if applicable) tests updated - [x] `CHANGELOG.md` updated (or `pr:no changelog` PR label added if not required) - [x] Documentation added (or `pr:no public docs` PR label added if not required) (will be done in #3762)
2024-01-30 13:07:19 -05:00
use_error_msg_as_fallback=False,
check_if_templated_value_is_truthy=False,
) -> typing.Union[str, None, bool]:
try:
templated_value = apply_jinja_template_to_alert_payload_and_labels(template, raw_request_data, labels)
return templated_value_is_truthy(templated_value) if check_if_templated_value_is_truthy else templated_value
except (JinjaTemplateError, JinjaTemplateWarning) as e:
fallback_msg = e.fallback_message
logger.warning(
f"{template_name} error on channel={alert_receive_channel.public_primary_key}: {fallback_msg}"
)
if use_error_msg_as_fallback:
return fallback_msg
elif check_if_templated_value_is_truthy:
return False
return None
@classmethod
def render_group_data(
cls,
alert_receive_channel: "AlertReceiveChannel",
raw_request_data: RawRequestData,
Support prescribed labels (#3848) # What this PR does **Cleanup label typing:** 1. LabelParam -> two separate types LabekKey and LabelValue 2. LabelData -> renamed to LabelPair. 3. LabelKeyData -> renamed to LabelOption Data is not giving any info about what this type represents. 4. Remove LabelsData and LabelsKeysData types. They are just list of types listed above and with new naming it feels obsolete. 5. ValueData removed. LabelPair is used instead. 6. Rework AlertGroupCustomLabel to use LabelKey type for key to make type system more consistent. Name model type AlertGroupCustomLabel**DB** and api type AlertGroupCustomLabel**API** to clearly distinguish them. **Split update_labels_cache into two tasks** update_label_option_cache and update_label_pairs_cache. Original task was expecting array of LabelsData (now it's LabelPair) OR one LabelKeyData ( now it's LabelOption). I believe having one function with two sp different argument types makes it more complicated for understanding. **Make OnCall backend support prescribed labels**. OnCall will sync and store "prescribed" field for key and values, so Label dropdown able to disable editing for certain labels. ## Which issue(s) this PR fixes ## Checklist - [x] Unit, integration, and e2e (if applicable) tests updated - [ ] Documentation added (or `pr:no public docs` PR label added if not required) - [x] `CHANGELOG.md` updated (or `pr:no changelog` PR label added if not required) --------- Co-authored-by: Maxim Mordasov <maxim.mordasov@grafana.com> Co-authored-by: Yulya Artyukhina <Ferril.darkdiver@gmail.com>
2024-02-20 14:42:51 +08:00
labels: typing.Optional[AlertLabels],
Support alert routing based on labels (#3778) # What this PR does This PR adds support for routing alerts based on labels. https://www.loom.com/share/4401de6e3c4945d5b8961fe43ee373c9 Additionally: - improve the typing around the `get_object` method that is inherited by [`PublicPrimaryKeyMixin.get_object`](https://github.com/grafana/oncall/blob/dev/engine/common/api_helpers/mixins.py#L153) in most of our models. `PublicPrimaryKeyMixin` is generic, so it can be more strongly typed when it is being subclassed, which results in better typing of the `get_object` method in child classes - I decided to do this because I started looking into this task via the [`AlertReceiveChannelView.send_demo_alert` method/endpoint](https://github.com/grafana/oncall/blob/dev/engine/apps/api/views/alert_receive_channel.py#L242). Within that method, `instance` is not typed because the inherited `get_object` method is not typed.. I digress 😄 - improve typing around `Alert.create` and `apps.integrations.tasks.create_alert` functions - make `Alert.render_group_data` more DRY by extracting some logic out into `Alert._apply_jinja_template_to_alert_payload_and_labels` - deduplicate the logic of `value.strip().lower() in ["1", "true", "ok"]` into a shared function, `common.jinja_templater.apply_jinja_template.templated_value_is_truthy` Closes https://github.com/grafana/oncall-private/issues/2490 ## Checklist - [x] Unit, integration, and e2e (if applicable) tests updated - [x] `CHANGELOG.md` updated (or `pr:no changelog` PR label added if not required) - [x] Documentation added (or `pr:no public docs` PR label added if not required) (will be done in #3762)
2024-01-30 13:07:19 -05:00
is_demo=False,
) -> "AlertGroup.GroupData":
`apps.get_model` -> `import` (#2619) # What this PR does Remove [`apps.get_model`](https://docs.djangoproject.com/en/3.2/ref/applications/#django.apps.apps.get_model) invocations and use inline `import` statements in places where models are imported within functions/methods to avoid circular imports. I believe `import` statements are more appropriate for most use cases as they allow for better static code analysis & formatting, and solve the issue of circular imports without being unnecessarily dynamic as `apps.get_model`. With `import` statements, it's possible to: - Jump to model definitions in most IDEs - Automatically sort inline imports with `isort` - Find import errors faster/easier (most IDEs highlight broken imports) - Have more consistency across regular & inline imports when importing models This PR also adds a flake8 rule to ban imports of `django.apps.apps`, so it's harder to use `apps.get_model` by mistake (it's possible to ignore this rule by using `# noqa: I251`). The rule is not enforced on directories with migration files, because `apps.get_model` is often used to get a historical state of a model, which is useful when writing migrations ([see this SO answer for more details](https://stackoverflow.com/a/37769213)). So `apps.get_model` is considered OK in migrations (even necessary in some cases). ## Checklist - [x] Unit, integration, and e2e (if applicable) tests updated - [x] Documentation added (or `pr:no public docs` PR label added if not required) - [x] `CHANGELOG.md` updated (or `pr:no changelog` PR label added if not required)
2023-07-25 10:43:23 +01:00
from apps.alerts.models import AlertGroup
template_manager = TemplateLoader()
is_resolve_signal = False
is_acknowledge_signal = False
Support alert routing based on labels (#3778) # What this PR does This PR adds support for routing alerts based on labels. https://www.loom.com/share/4401de6e3c4945d5b8961fe43ee373c9 Additionally: - improve the typing around the `get_object` method that is inherited by [`PublicPrimaryKeyMixin.get_object`](https://github.com/grafana/oncall/blob/dev/engine/common/api_helpers/mixins.py#L153) in most of our models. `PublicPrimaryKeyMixin` is generic, so it can be more strongly typed when it is being subclassed, which results in better typing of the `get_object` method in child classes - I decided to do this because I started looking into this task via the [`AlertReceiveChannelView.send_demo_alert` method/endpoint](https://github.com/grafana/oncall/blob/dev/engine/apps/api/views/alert_receive_channel.py#L242). Within that method, `instance` is not typed because the inherited `get_object` method is not typed.. I digress 😄 - improve typing around `Alert.create` and `apps.integrations.tasks.create_alert` functions - make `Alert.render_group_data` more DRY by extracting some logic out into `Alert._apply_jinja_template_to_alert_payload_and_labels` - deduplicate the logic of `value.strip().lower() in ["1", "true", "ok"]` into a shared function, `common.jinja_templater.apply_jinja_template.templated_value_is_truthy` Closes https://github.com/grafana/oncall-private/issues/2490 ## Checklist - [x] Unit, integration, and e2e (if applicable) tests updated - [x] `CHANGELOG.md` updated (or `pr:no changelog` PR label added if not required) - [x] Documentation added (or `pr:no public docs` PR label added if not required) (will be done in #3762)
2024-01-30 13:07:19 -05:00
group_distinction: typing.Optional[str] = None
web_title_cache: typing.Optional[str] = None
# set web_title_cache to web title to allow alert group searching based on web_title_cache
Support alert routing based on labels (#3778) # What this PR does This PR adds support for routing alerts based on labels. https://www.loom.com/share/4401de6e3c4945d5b8961fe43ee373c9 Additionally: - improve the typing around the `get_object` method that is inherited by [`PublicPrimaryKeyMixin.get_object`](https://github.com/grafana/oncall/blob/dev/engine/common/api_helpers/mixins.py#L153) in most of our models. `PublicPrimaryKeyMixin` is generic, so it can be more strongly typed when it is being subclassed, which results in better typing of the `get_object` method in child classes - I decided to do this because I started looking into this task via the [`AlertReceiveChannelView.send_demo_alert` method/endpoint](https://github.com/grafana/oncall/blob/dev/engine/apps/api/views/alert_receive_channel.py#L242). Within that method, `instance` is not typed because the inherited `get_object` method is not typed.. I digress 😄 - improve typing around `Alert.create` and `apps.integrations.tasks.create_alert` functions - make `Alert.render_group_data` more DRY by extracting some logic out into `Alert._apply_jinja_template_to_alert_payload_and_labels` - deduplicate the logic of `value.strip().lower() in ["1", "true", "ok"]` into a shared function, `common.jinja_templater.apply_jinja_template.templated_value_is_truthy` Closes https://github.com/grafana/oncall-private/issues/2490 ## Checklist - [x] Unit, integration, and e2e (if applicable) tests updated - [x] `CHANGELOG.md` updated (or `pr:no changelog` PR label added if not required) - [x] Documentation added (or `pr:no public docs` PR label added if not required) (will be done in #3762)
2024-01-30 13:07:19 -05:00
if (
web_title_template := template_manager.get_attr_template("title", alert_receive_channel, render_for="web")
) is not None:
web_title_cache = cls._apply_jinja_template_to_alert_payload_and_labels(
web_title_template,
"web_title_cache",
alert_receive_channel,
raw_request_data,
labels,
use_error_msg_as_fallback=True,
)
if (
grouping_id_template := template_manager.get_attr_template("grouping_id", alert_receive_channel)
) is not None:
group_distinction = cls._apply_jinja_template_to_alert_payload_and_labels(
grouping_id_template, "grouping_id_template", alert_receive_channel, raw_request_data, labels
)
# Insert random uuid to prevent grouping of demo alerts or alerts with group_distinction=None
if is_demo or not group_distinction:
group_distinction = cls.insert_random_uuid(group_distinction)
if group_distinction is not None:
group_distinction = hashlib.md5(str(group_distinction).encode()).hexdigest()
Support alert routing based on labels (#3778) # What this PR does This PR adds support for routing alerts based on labels. https://www.loom.com/share/4401de6e3c4945d5b8961fe43ee373c9 Additionally: - improve the typing around the `get_object` method that is inherited by [`PublicPrimaryKeyMixin.get_object`](https://github.com/grafana/oncall/blob/dev/engine/common/api_helpers/mixins.py#L153) in most of our models. `PublicPrimaryKeyMixin` is generic, so it can be more strongly typed when it is being subclassed, which results in better typing of the `get_object` method in child classes - I decided to do this because I started looking into this task via the [`AlertReceiveChannelView.send_demo_alert` method/endpoint](https://github.com/grafana/oncall/blob/dev/engine/apps/api/views/alert_receive_channel.py#L242). Within that method, `instance` is not typed because the inherited `get_object` method is not typed.. I digress 😄 - improve typing around `Alert.create` and `apps.integrations.tasks.create_alert` functions - make `Alert.render_group_data` more DRY by extracting some logic out into `Alert._apply_jinja_template_to_alert_payload_and_labels` - deduplicate the logic of `value.strip().lower() in ["1", "true", "ok"]` into a shared function, `common.jinja_templater.apply_jinja_template.templated_value_is_truthy` Closes https://github.com/grafana/oncall-private/issues/2490 ## Checklist - [x] Unit, integration, and e2e (if applicable) tests updated - [x] `CHANGELOG.md` updated (or `pr:no changelog` PR label added if not required) - [x] Documentation added (or `pr:no public docs` PR label added if not required) (will be done in #3762)
2024-01-30 13:07:19 -05:00
if (
resolve_condition_template := template_manager.get_attr_template("resolve_condition", alert_receive_channel)
) is not None:
is_resolve_signal = cls._apply_jinja_template_to_alert_payload_and_labels(
resolve_condition_template,
"resolve_condition_template",
alert_receive_channel,
raw_request_data,
labels,
check_if_templated_value_is_truthy=True,
)
if (
acknowledge_condition_template := template_manager.get_attr_template(
"acknowledge_condition", alert_receive_channel
)
) is not None:
is_acknowledge_signal = cls._apply_jinja_template_to_alert_payload_and_labels(
acknowledge_condition_template,
"acknowledge_condition_template",
alert_receive_channel,
raw_request_data,
labels,
check_if_templated_value_is_truthy=True,
)
return AlertGroup.GroupData(
is_resolve_signal=is_resolve_signal,
is_acknowledge_signal=is_acknowledge_signal,
group_distinction=group_distinction,
web_title_cache=web_title_cache,
)
@staticmethod
Support alert routing based on labels (#3778) # What this PR does This PR adds support for routing alerts based on labels. https://www.loom.com/share/4401de6e3c4945d5b8961fe43ee373c9 Additionally: - improve the typing around the `get_object` method that is inherited by [`PublicPrimaryKeyMixin.get_object`](https://github.com/grafana/oncall/blob/dev/engine/common/api_helpers/mixins.py#L153) in most of our models. `PublicPrimaryKeyMixin` is generic, so it can be more strongly typed when it is being subclassed, which results in better typing of the `get_object` method in child classes - I decided to do this because I started looking into this task via the [`AlertReceiveChannelView.send_demo_alert` method/endpoint](https://github.com/grafana/oncall/blob/dev/engine/apps/api/views/alert_receive_channel.py#L242). Within that method, `instance` is not typed because the inherited `get_object` method is not typed.. I digress 😄 - improve typing around `Alert.create` and `apps.integrations.tasks.create_alert` functions - make `Alert.render_group_data` more DRY by extracting some logic out into `Alert._apply_jinja_template_to_alert_payload_and_labels` - deduplicate the logic of `value.strip().lower() in ["1", "true", "ok"]` into a shared function, `common.jinja_templater.apply_jinja_template.templated_value_is_truthy` Closes https://github.com/grafana/oncall-private/issues/2490 ## Checklist - [x] Unit, integration, and e2e (if applicable) tests updated - [x] `CHANGELOG.md` updated (or `pr:no changelog` PR label added if not required) - [x] Documentation added (or `pr:no public docs` PR label added if not required) (will be done in #3762)
2024-01-30 13:07:19 -05:00
def insert_random_uuid(distinction: typing.Optional[str]) -> str:
if distinction is not None:
distinction += str(uuid4())
else:
distinction = str(uuid4())
return distinction