4. Cost monitoring
You'll complete a short day-zero cost path right after infra: see spend in the account console, optionally add dashboards in a workspace, tag workloads, and set budgets.
Prereqs: Infra setup
Set this up before you have anything to monitor
Turn on cost visibility now, while the account is small. Teams that wait pay for it twice: a surprise invoice when usage scales past whoever was supposed to be watching, and an awkward silence when finance asks who spent what and nobody can say. Both go away if consumption is already visible when you have a handful of workspaces and known owners.
The other reason to do it early is tags. They only apply to usage created after you set them, so every week you wait is a week of spend you can never attribute. Retrofitting attribution is not a thing.
Starter Journey Progress
How to think about system tables
System tables are read-only billing and operations tables in the system catalog, for example system.billing.usage and system.billing.list_prices. Every dashboard you import and every cost query you write later reads from them. You do not need to learn the schema on day zero. You need to know these tables are the source of truth for consumption and list-price dollars, and everything else in this section is a view on top of them.
For depth, use the official references:
- System tables reference
- Billable usage system table reference
- Monitor costs using system tables
- Top 10 queries to use with System Tables
Recommended order
| Step | Topic | Why this slot |
|---|---|---|
| 1 | Import usage dashboard | Fastest visibility, straight from the account console, no SQL. |
| 2 | Additional dashboards | Optional packaged dashboards on your catalog and schema. |
| 3 | Tags and attribution | Forward-only labels for team and project rollups. |
| 4 | Budget alerts | Account-level monthly monitors with email. |
Next
- Do next: Import usage dashboard
- Learn why: Account Console foundations
- Reference: Cost management tools on Databricks