Interface inventory: the text here is the text on screen.
Where it is: at the /datalake address. Since 09/26/2026 the screen no longer appears
in the menu — the address still works, and that's how you get here (or through the
documentation links). For day-to-day work, the path is
Data → Console.
Title Your datalake. The sentence at the top sums up the screen: "This is where your data is stored, in three layers: as it arrived, already cleaned, and ready for dashboards."
The three layers#
With the definition the screen itself uses:
| Layer | What it is |
|---|---|
| Bronze | An exact copy of what came from your sources — nothing corrected. |
| Silver | Clean, standardized data, ready to combine. |
| Gold | Metrics already aggregated, ready for dashboards. |
Dashboards usually read from Gold. What moves data from one layer to another is the pipeline.
Create the datalake (first time only)#
1. Open the screen
Go to the /datalake address. If it doesn't exist yet, the screen shows Your datalake hasn't been created yet.
2. Click Provision datalake
The button creates the three datasets (bronze, silver, and gold) for your environment. The step-by-step appears live, one item at a time.
3. Check the status
Under Datalake status, the badge changes to Provisioned.
4. Bring in data
Use Connect a source to register access to your system and then build a pipeline.
If the badge already shows Provisioned, the datalake exists and this step isn't needed.
What each button does#
| You want to | Click |
|---|---|
| Create the datalake | Provision datalake |
| Redo the creation steps | Reprovision datalake (appears after it exists) |
| Look at the tables with SQL | Explore your data (opens the Console) |
| Bring in data from a system | Connect a source (opens Connections) |
Reprovisioning doesn't delete anything
The How your datalake was created (and how to redo the steps) section stays collapsed and is only for checking. Each step recreates whatever is missing and keeps what already exists: "No table and no data is deleted from here."
What else the screen shows#
- Datalake status — Provisioned · Pending · Provisioning error.
- Your tables — Dataset · Table · Rows. Before the first load: "No tables yet — they show up here after a source's first load."
- Where your data lives — "Three different facts, which don't have to match.": Dataset location, Where queries are processed, and Where the app runs. Your environment ID is shown only in part — "Use it when contacting support".
- Isolation and billing — "Each customer's data lives in its own datasets, with its own permissions. Your queries never reach another customer's tables."
Back to the app#
- Datalake: Your datalake
- Console: Data → Console