Is there a cost for Databricks Workflow notifications or are they free?

[MageshS]

I couldn't find pricing information for Workflow/Job notifications anywhere in the Documentation or on the Pricing Page.

Is sending notifications for job run events (start, success, failure) via email, Slack, webhook, MS Teams, or PagerDuty a free feature, or is there some cost (DBU or otherwise) associated with it? Could someone confirm with an official source or doc reference? Thanks!

[balajij8]

The notification mechanisms such as routing alerts for start, success or failure events via email, Slack, MS Teams, PagerDuty or generic webhooks are included as part of the native orchestration feature set at no additional charge currently.

Databricks bills strictly for the compute resources (DBUs and underlying cloud infrastructure) that the job tasks actually consume while executing. The control plane handles the scheduling, orchestration and firing off of these alerts. The official documentation omits any mention of billing because there is none currently. 

While Databricks won't charge you to send the notification, downstream services (PagerDuty or custom webhook listeners) might have their own ingestion limits or API costs. You can audit the billing tables with the jobs system tables in the system.lakeflow schema to find what a job is costing you.

[Louis_Frolio]

Hi @MageshS ,

Good question, and you are not missing anything on the Pricing Page. There is no separate line item for job notifications because they are not billed as their own feature.

@balajij8 has this right. Notifications for job run events (start, success, failure, and the duration-threshold warning) are part of the native orchestration in Lakeflow Jobs. Sending them costs nothing extra, whether you route to email, Slack, MS Teams, PagerDuty, or a generic webhook. There is no notification-specific SKU in the public docs.

Here is the mental model I use. Databricks bills you for the compute your tasks consume while they run, measured in DBUs, plus the underlying cloud infrastructure. On serverless it shows up as pay-as-you-go compute. The control plane handles the scheduling, the orchestration, and the firing of alerts, and that piece is included. So the cost of a job is the cost of the compute it runs on, not the act of telling you it finished.

Two things to keep in the back of your mind:

1. Downstream services can have their own costs. This was @balajij8's point and it is a good one. Databricks will not charge you to send the notification, but PagerDuty, a custom webhook listener, or an email gateway on your side may carry its own ingestion limits or API charges. That sits with the receiving system, not with Databricks.
2. If you want to see what a job actually costs, the system tables are the place to look. One small refinement on the earlier reply: the dollar and DBU figures live in system.billing.usage (with SKU pricing in system.billing.list_prices), while the job and run metadata live in the system.lakeflow schema (system.lakeflow.jobs and system.lakeflow.job_run_timeline). You attribute cost to a specific job by joining the billing usage to the run timeline, and the docs include a ready-made query for exactly that.

A few references you can cite:

• Add notifications on a job: https://docs.databricks.com/aws/en/jobs/notifications
• Manage notification destinations (Slack, Teams, PagerDuty, webhook): https://docs.databricks.com/aws/en/admin/workspace-settings/notification-destinations
• Lakeflow Jobs product page (serverless pay-as-you-go pricing): https://www.databricks.com/product/data-engineering/lakeflow-jobs
• Monitor job costs and performance with system tables: https://docs.databricks.com/aws/en/admin/system-tables/jobs-cost
• Jobs system table reference: https://docs.databricks.com/aws/en/admin/system-tables/jobs

One honest caveat: I am describing current behavior. Databricks has not billed for notifications as far as I know, but pricing can change over time, so treat the system tables as your source of truth for what any given job is really costing you.

Regards, Louis