dagster-apprise 0.0.3

Get notified when your Dagster runs succeed, fail, or change status across 70+ services including Slack, Discord, Pushover, Telegram, and more.

dagster-apprise

A dagster module that provides integration with Apprise for sending notifications across 70+ services including Slack, Discord, Pushover, Telegram, email, and more.

Installation

The dagster_apprise module is a PyPI package - use whichever python environment manager you'd prefer, although we default to using uv.

uv venv
source .venv/bin/activate
uv pip install dagster-apprise

Example Usage

import dagster as dg
from dagster_apprise import apprise_notifications, AppriseNotificationsConfig

# Configure notifications
config = AppriseNotificationsConfig(
    urls=["discord://webhook_id/webhook_token"], # Discord notification, details should be secure
    events=["SUCCESS", "FAILURE"],  # What to notify about
    include_jobs=["*"],  # All jobs
)

# Add to your Dagster definitions
defs = dg.Definitions(
    assets=[your_assets],
    jobs=[your_jobs],
    **apprise_notifications(config).to_dict()
)

Configuration Options

Full List

See the Apprise documentation for all supported services.

AppriseNotificationsConfig

Option	Type	Default	Description
urls	list[str]	[]	List of Apprise notification URLs
config_file	str	None	Path to Apprise config file
base_url	str	None	Base URL for Dagster UI links
title_prefix	str	"Dagster"	Prefix for notification titles
events	list[str]	["FAILURE"]	Events to notify about
include_jobs	list[str]	["*"]	Job patterns to include
exclude_jobs	list[str]	[]	Job patterns to exclude

Supported Events

• SUCCESS: Job completed successfully
• FAILURE: Job failed
• CANCELED: Job was canceled
• RUNNING: Job started running

Job Filtering

Use wildcard patterns to control which jobs trigger notifications:

config = AppriseNotificationsConfig(
    urls=["mailto://user:[e-mail]"],
    include_jobs=["prod_*", "critical_*"],  # Only production and critical jobs
    exclude_jobs=["*_test", "*_dev"],       # Exclude test and dev jobs
    events=["SUCCESS", "FAILURE"],
)

Advanced Usage

Op-level Notifications with Hooks

from dagster_apprise import apprise_on_failure, apprise_on_success
import dagster as dg

@apprise_on_failure(urls=["mailto://user:[e-mail]"])
@dg.job
def critical_job():
    my_op()

# Or apply to specific ops
@dg.job
def selective_notifications():
    my_op.with_hooks({apprise_on_failure(urls=["mailto://user:[e-mail]"])})()

Using AppriseResource in Ops

import dagster as dg
from dagster_apprise import AppriseResource

@dg.op
def notify_op(apprise: AppriseResource):
    apprise.notify(title="Custom Alert", body="processing complete")

Using with Environment Variables

import os
from dagster_apprise import AppriseNotificationsConfig

# Load from environment
config = AppriseNotificationsConfig(
    urls=[os.getenv("GMAIL_URL")],
    base_url=os.getenv("DAGSTER_BASE_URL", "http://localhost:3000"),
    events=["SUCCESS", "FAILURE"],
)

YAML Configuration

Create defs.yaml:

type: dagster_apprise.apprise_notifications
attributes:
  urls:
    - "pover://user@token"
    - "https://hooks.slack.com/services/YOUR/SLACK/WEBHOOK"
  base_url: "http://localhost:3000"
  events: ["SUCCESS", "FAILURE"]
  include_jobs: ["*"]
  exclude_jobs: ["*_test"]

Multiple Notification Channels

config = AppriseNotificationsConfig(
    urls=[
        "pover://user@token",  # Pushover for immediate testing alerts
        "https://hooks.slack.com/services/YOUR/SLACK/WEBHOOK",  # Slack for team
        "mailto://user:[e-mail]",  # Email for backup
    ],
    events=["SUCCESS", "FAILURE"],
)

Different Notifications for Different Jobs

# Critical jobs - notify on all events
critical_config = AppriseNotificationsConfig(
    urls=["mailto://user:[e-mail]"],
    include_jobs=["critical_*"],
    events=["SUCCESS", "FAILURE", "CANCELED", "RUNNING"],
)

# Regular jobs - only notify on failures
regular_config = AppriseNotificationsConfig(
    urls=["https://hooks.slack.com/services/YOUR/SLACK/WEBHOOK"],
    include_jobs=["*"],
    exclude_jobs=["critical_*"],
    events=["FAILURE"],
)

# Combine both
defs = dg.Definitions(
    assets=[your_assets],
    jobs=[your_jobs],
    resources={
        **apprise_notifications(critical_config).resources,
        **apprise_notifications(regular_config).resources,
    },
    sensors=[
        *apprise_notifications(critical_config).sensors,
        *apprise_notifications(regular_config).sensors,
    ],
)

Custom Run Failure Sensor

from dagster_apprise import make_apprise_on_run_failure_sensor

apprise_on_failure = make_apprise_on_run_failure_sensor(
    urls=["mailto://user:[e-mail]"],
    monitored_jobs=["prod_*"]
)

defs = dg.Definitions(sensors=[apprise_on_failure])

Integration with Dagster Logs

@dg.op
def my_op(context):
    context.log.info("This appears in Dagster logs")
    error_condition = False
    if error_condition:
        context.resources.apprise.notify(
            title="Critical Error",
            body=f"Job {context.job_name} encountered an error",
        )

Dagster+ Users

If you're using Dagster+, you already have built-in alerting for Slack, Microsoft Teams, PagerDuty, and email. For OSS users, try dagster-apprise for an alternative.

Notes on Base URL

• You can set base_url or webserver_base_url (alias) to include deep links to runs in notifications.

Development

The Makefile provides the tools required to test and lint your local installation.

make test
make ruff
make check