YALineo
EN
Install

The map, cost and budget of your Databricks platform.

For data teams on Databricks: the YALineo extension draws your lineage as a metro map, gives the cost of each table and prices a new job before it is created.

Free today, no account. Installs in Chrome, Edge and Firefox through the "Install" button.

Data and lineage Platform mapClick a station to open its sheet: its map, its cost and its budget.

Inventory lineTable · silver

inventory_stock_level

Map
1 direct upstream, 2 direct downstreamAlong the whole chain: 2 upstream, 8 downstream. Owner: data-eng-retail.
Cost
$471.44 over 90 days
precision: compute time

A run that writes several tables: its cost is shared by the compute time of the queries writing each.

33 reads observed, $14.29 per read.
Budget
$159.33 a monthAt the measured pace, scaled to one month.

Sales lineTable · gold

fact_sales

Map
5 direct upstream, 5 direct downstreamAlong the whole chain: 17 upstream, 9 downstream. Owner: data-eng-retail.
Cost
$423.87 over 90 days
precision: even split

A run that writes several tables without a compute time for all of them: equal shares.

131 reads observed, $3.24 per read.
Budget
$143.25 a monthAt the measured pace, scaled to one month.

E-commerce lineTable · gold

fact_ecommerce_cart

Map
1 direct upstream, 0 direct downstreamAlong the whole chain: 3 upstream, 0 downstream. Owner: data-eng-retail.
Cost
$423.87 over 90 days
precision: even split

A run that writes several tables without a compute time for all of them: equal shares.

No read observed over the period.
Budget
$143.25 a monthRefreshed with no read observed. Check with data-eng-retail before stopping anything.

Sales lineDashboard · consumption

Store operations

Map
2 direct upstream, 0 direct downstream21 tables refreshed for it along the whole chain.
Cost
$4,642.76 over 90 days to keep its data up to dateThe sum of the refreshes of its chain, table by table.
Budget
$1,569.08 a monthAt the measured pace, scaled to one month.

Inventory linePlanned table · gold

New gold table

Map
Not built yet. No measured map contains it.
Cost
No measured cost, since nothing has run yet.
Budget
$69.90 per month, estimatedAdvanced pipeline, 2 workers, 2 runs a day of 30 minutes. Spot, the cheapest: $59.06. Azure public prices, East US, collected on October 15, 2026. An estimate, never a measure.

Sales lineExternal source · source

pos/receipt

Map
It feeds 21 assets, all the way to the dashboards.
Cost
No cost of its own in the Databricks bill. YALineo counts the workload that reads it in the table that workload writes.
Budget
Nothing to plan for the source itself.

Sales lineStreaming table · bronze

sales_receipt

Map
1 direct upstream, 1 direct downstreamAlong the whole chain: 1 upstream, 20 downstream. Owner: data-eng-retail.
Cost
$134.38 over 90 days
precision: even split

A run that writes several tables without a compute time for all of them: equal shares.

35 reads observed, $3.84 per read.
Budget
$45.41 a monthAt the measured pace, scaled to one month.

Sales lineTable · silver

sales_receipt

Map
1 direct upstream, 3 direct downstreamAlong the whole chain: 2 upstream, 19 downstream. Owner: data-eng-retail.
Cost
$109.79 over 90 days
precision: compute time

A run that writes several tables: its cost is shared by the compute time of the queries writing each.

106 reads observed, $1.04 per read.
Budget
$37.10 a monthAt the measured pace, scaled to one month.

Sales lineMaterialized view · platinum

store_performance

Map
3 direct upstream, 1 direct downstreamAlong the whole chain: 19 upstream, 1 downstream. Owner: bi-team.
Cost
$262.00 over 90 days
precision: even split

A run that writes several tables without a compute time for all of them: equal shares.

8 reads observed, $32.75 per read.
Budget
$88.55 a monthAt the measured pace, scaled to one month.

Stores lineExternal source · source

storeops/location

Map
It feeds 20 assets, all the way to the dashboards.
Cost
No cost of its own in the Databricks bill. YALineo counts the workload that reads it in the table that workload writes.
Budget
Nothing to plan for the source itself.

Stores lineStreaming table · bronze

stores_location

Map
1 direct upstream, 1 direct downstreamAlong the whole chain: 1 upstream, 19 downstream. Owner: data-eng-stores.
Cost
$134.38 over 90 days
precision: even split

A run that writes several tables without a compute time for all of them: equal shares.

32 reads observed, $4.20 per read.
Budget
$45.41 a monthAt the measured pace, scaled to one month.

Stores lineTable · silver

stores_location

Map
1 direct upstream, 1 direct downstreamAlong the whole chain: 2 upstream, 18 downstream. Owner: data-eng-stores.
Cost
$114.41 over 90 days
precision: compute time

A run that writes several tables: its cost is shared by the compute time of the queries writing each.

33 reads observed, $3.47 per read.
Budget
$38.67 a monthAt the measured pace, scaled to one month.

Stores lineTable · gold

dim_store

Map
2 direct upstream, 7 direct downstreamAlong the whole chain: 6 upstream, 17 downstream. Owner: data-eng-retail.
Cost
$423.87 over 90 days
precision: even split

A run that writes several tables without a compute time for all of them: equal shares.

178 reads observed, $2.38 per read.
Budget
$143.25 a monthAt the measured pace, scaled to one month.

Stores lineMaterialized view · platinum

store_shelf_availability

Map
2 direct upstream, 1 direct downstreamAlong the whole chain: 14 upstream, 1 downstream. Owner: bi-team.
Cost
$122.00 over 90 days
precision: even split

A run that writes several tables without a compute time for all of them: equal shares.

27 reads observed, $4.52 per read.
Budget
$41.23 a monthAt the measured pace, scaled to one month.

Stores lineDashboard · consumption

Shelf availability

Map
1 direct upstream, 0 direct downstream11 tables refreshed for it along the whole chain.
Cost
$1,920.42 over 90 days to keep its data up to dateThe sum of the refreshes of its chain, table by table.
Budget
$649.03 a monthAt the measured pace, scaled to one month.

Inventory lineExternal source · source

wms/stock_level

Map
It feeds 10 assets, all the way to the dashboards.
Cost
No cost of its own in the Databricks bill. YALineo counts the workload that reads it in the table that workload writes.
Budget
Nothing to plan for the source itself.

Inventory lineStreaming table · bronze

inventory_stock_level

Map
1 direct upstream, 1 direct downstreamAlong the whole chain: 1 upstream, 9 downstream. Owner: data-eng-retail.
Cost
$134.38 over 90 days
precision: even split

A run that writes several tables without a compute time for all of them: equal shares.

28 reads observed, $4.80 per read.
Budget
$45.41 a monthAt the measured pace, scaled to one month.

Inventory lineTable · gold

fact_inventory

Map
5 direct upstream, 3 direct downstreamAlong the whole chain: 20 upstream, 7 downstream. Owner: data-eng-retail.
Cost
$423.87 over 90 days
precision: even split

A run that writes several tables without a compute time for all of them: equal shares.

87 reads observed, $4.87 per read.
Budget
$143.25 a monthAt the measured pace, scaled to one month.

Inventory lineMaterialized view · platinum

inventory_turnover

Map
2 direct upstream, 2 direct downstreamAlong the whole chain: 21 upstream, 2 downstream. Owner: bi-team.
Cost
$122.00 over 90 days
precision: even split

A run that writes several tables without a compute time for all of them: equal shares.

32 reads observed, $3.81 per read.
Budget
$41.23 a monthAt the measured pace, scaled to one month.

Inventory lineDashboard · consumption

Inventory turnover

Map
1 direct upstream, 0 direct downstream16 tables refreshed for it along the whole chain.
Cost
$3,209.35 over 90 days to keep its data up to dateThe sum of the refreshes of its chain, table by table.
Budget
$1,084.64 a monthAt the measured pace, scaled to one month.

E-commerce lineExternal source · source

webshop/cart

Map
It feeds 3 assets, all the way to the dashboards.
Cost
No cost of its own in the Databricks bill. YALineo counts the workload that reads it in the table that workload writes.
Budget
Nothing to plan for the source itself.

E-commerce lineStreaming table · bronze

ecommerce_cart

Map
1 direct upstream, 1 direct downstreamAlong the whole chain: 1 upstream, 2 downstream. Owner: data-eng-digital.
Cost
$134.38 over 90 days
precision: even split

A run that writes several tables without a compute time for all of them: equal shares.

35 reads observed, $3.84 per read.
Budget
$45.41 a monthAt the measured pace, scaled to one month.

E-commerce lineTable · silver

ecommerce_cart

Map
1 direct upstream, 1 direct downstreamAlong the whole chain: 2 upstream, 1 downstream. Owner: data-eng-digital.
Cost
$84.33 over 90 days
precision: compute time

A run that writes several tables: its cost is shared by the compute time of the queries writing each.

21 reads observed, $4.02 per read.
Budget
$28.50 a monthAt the measured pace, scaled to one month.

  • Sales
  • Stores
  • Inventory
  • E-commerce
How each figure is calculated
The extension's example platform, fictional data: a retail chain, over 90 days. The YALineo engine computes every figure.
  • CloudsAzure, AWS and GCP
  • BrowsersChrome, Edge and Firefox
  • LanguagesFrench and English

YALineo in one minute.

Four people, four questions, and each side's answer, on the real interface of the extension.

Example platform, fictional data.
Read the film as text

The film has no voice. Music plays under it, and all its meaning is in the words on screen, given here scene by scene.

  1. Four people each ask a question. The head of store operations: "Where's this number from?" The data engineer: "What will this change break?" The financial controller: "Why is the bill rising?" The head of data: "What will the new job cost?" Twenty more questions pile up, and this month's bill climbs to $4,221.
  2. Everything stops on one line: "The same questions, every week."
  3. The four questions line up on the left. In the workspace, the YALineo chip lights up and the studio opens on "One extension. Three sides." The sides separate: Lineage, FinOps and Calculator.
  4. First question, with Lineage. The transit map of the platform, a search for the web orders table, then its line from upstream to downstream and its sheet. Answer: "The source and downstream of every table."
  5. Second question, with impact analysis. A change to the web orders table, and what it hits, from the most to the least serious. Answer: "12 assets hit, 4 owners to notify."
  6. Third question, with FinOps. The spend of the period, the tables refreshed with no read observed, then the tables ranked by refresh cost. Answer: "Every table has its price."
  7. Fourth question, with the Calculator. Loading 40 million rows a night takes 72 minutes. Its price: $69.96 a month on spot, or $839.52 a year. The other ways to pay, per month: $96.00 pay as you go, $541.40 reserved 1 year, $368.21 reserved 3 years. Answer: "The price, before you build."
  8. All four questions have their answer. "Read-only. No server. Right inside Databricks." The cards gather into the YALineo mark, then come the name and the address, yalineo.com.

Three questions come back every week.

Each side answers one of them, with the system tables of your workspace and your permissions.

Lineage · The map

What depends on this table?

Before changing it, you need to know which jobs, tables and dashboards break behind it. Catalog Explorer shows one level at a time.

  • The platform map
  • The dependencies of an asset
  • The sheet of an asset
  • The impact of a change
  • Why no link?
  • The catalog and the glossary
  • To review
  • Pull request preflight
See the Lineage screens
Screenshot of YALineo: the dependencies of fact_inventory, from its external sources to the dashboards, zone by zoneScreenshot of YALineo: the sheet of the fact_inventory table, with its line, its layer, its freshness, its direct neighbors and its whole chain
The sheet of the fact_inventory table: its line, its freshness, its direct neighbors, then its whole chain. Behind it, its dependencies, from the external sources to the dashboards. Example platform, fictional data.

FinOps · The cost

What does this table cost, and who reads it?

The bill says which job consumes. It does not say which table it refreshes, or whether anyone still reads it. The costs shown are those of Databricks billing (DBUs and other Databricks units), at list price or with your discount; the classic compute machines, storage and network billed by your cloud are not in them.

  • The cost overview
  • The cost of every table
  • Catalogs and schemas
  • Tables and views, each in its place
  • Tables with no read observed
  • By workload type
  • Cost analysis
  • Projections and the cost map
See the FinOps screens
Screenshot of YALineo: the cost map, with the schemas of the platform, the number of their tables, what they cost and, as a halo, their share of the costScreenshot of YALineo: the list of the tables that cost most to refresh, each with its cost per month
The tables that cost most to refresh, each with its cost per month. Behind it, the cost map: the schemas, what their tables cost, and as a halo the share of each. Example platform, fictional data.

Calculator · The budget

What will the next workload cost?

Before a new job, more data or another cloud, you need a budget. Public calculators ignore what your workloads already consume.

  • Price it before you build it
  • Cost a data project
  • Size and tune
  • Genie and AI agents
  • The four ways to pay
  • Start from a real job
  • Public prices and your account's prices
  • Compare and share
  • An assistant, if you turn it on
See the Calculator screens
Screenshot of YALineo: one workload of the estimate, each value of its calculation, then its calculation step by stepScreenshot of YALineo: the total of an estimate per month and per year, its pricing state, and its share in DBUs and in machines
The total of an estimate, per month and per year, split between the Databricks DBUs and the cloud machines. Beside it, one of its workloads and its calculation, step by step. Example platform, fictional data.

Privacy

By default, your metadata only travels between your workspace and your browser.

  • Read-only

    YALineo only sends SELECT statements. It never writes to Unity Catalog.

  • No server

    No account, no telemetry. The extension only talks to Databricks domains.

  • Your permissions, no more

    Every query goes out with your session or your token, so you only see what Databricks already shows you.

  • Map erased by default

    The map is erased when the browser closes (a map too big for session memory, at the browser's next start). One click erases it now.

For a security review, the source code is available on request.

Three steps to get started.

  1. Install

    From your browser's store: Chrome, Edge or Firefox.

  2. Connect

    With the Databricks session already open, or with a personal token.

  3. Read the map

    Pick a SQL warehouse. YALineo loads the map and the costs read only, with your permissions.

Before you start

You need a Databricks workspace with Unity Catalog and a SQL warehouse. The full map reads system.access; the FinOps side reads system.billing. These system tables are often admin-only: YALineo provides the grants script, block by block, to hand to an admin. Without them, analyst mode draws the map with the lineage API. Without a workspace or permissions, "See the fictional example" opens an example platform with fictional data: its map, its costs and the Calculator.

Step by step: the tutorials

Free today, no account.

Installs on a computer, in Chrome, Edge or Firefox. On a phone: request a demo or send yourself this page's address.

The browser asks for access to the Databricks sites: they are the only ones YALineo can read, and it only sends them reads. Security and evidence

See your platform in 30 minutes.

On a video call, on your own data read only or on the example platform. On your data, you leave with the map of your platform and the cost of your tables.

Request a demo

The button opens your mail app, Outlook for example, with a message ready to complete.

On the agenda

  1. The whole map, from sources to reports
  2. What each table, job and dashboard costs
  3. The tables refreshed with no read observed
  4. The price of your next project
  5. Your security and permission questions
Which permissions are needed in Databricks?

For the full map, read access to system.access.table_lineage. For the FinOps side, read access to system.billing.usage and system.billing.list_prices; system.lakeflow gives the names of the jobs and their runs, and system.query.history the reads on SQL warehouses. These system tables are often admin-only, and YALineo provides a ready-made grants script, block by block. Without them, analyst mode draws the map with the Unity Catalog lineage API, table by table.

Does my data leave my network?

Not by default. There is no YALineo server: queries go from your browser to your workspace, and the answers stay in your browser. Three cases are the exceptions. The Calculator's assistant, off by default, sends your description to the serving endpoint you name, and Databricks passes it on to the provider if that endpoint serves an external model. The pull-request preflight, if you install it in your continuous integration, writes the tables touched, the dashboards that depend on them and their owners in your pull requests. If your browser's sync is on, turned on by you or by your organization, the addresses of the extension's pages (the studio with its followed table and filters; the settings) go with it, like any address. The rest leaves only by your click, for example an email to support that you read first, an email draft you address to anyone you like, a copy to the clipboard or files you export.

How can we check YALineo stays read only?

The studio's "Diagnostics" page checks it right now, on your device. It makes sure the read-only guard refuses a write and lets a read through, and that the extension's network permissions only go to the Databricks domains. It also checks the last reading of the map, the errors of the last 24 hours and the age of the map. When everything passes, a green badge shows "YALineo is secure and healthy". Otherwise it turns orange or red and names the failing check.

Are the costs exact?

They come from Databricks billing, at list price, and you set a negotiated discount in the settings. Each amount shows its nature (observed, attributed, estimated or potential) and its precision: billed to the table, tied to the job run that wrote it, split hour by hour by the compute time of a SQL warehouse, or split over the period when no finer data exists. Next to a zero, missing or partial cost, "Why?" gives the measured reason. What links to no table stays visible. The FinOps side does not include the cloud bill (machines, storage). The "Method" page gives each rule and its official source.

Azure, AWS or GCP?

All three. YALineo works on any Databricks workspace with Unity Catalog, in light and dark themes.

Does YALineo go through an AI?

Only if you decide so. The glossary uses none to extract or import: it reads, in the browser, what Unity Catalog already says or your file. For tables without a comment, YALineo can write an ai_query statement you run yourself, and an estimate of the calls and tokens shows first. The Calculator's assistant is off by default. If you turn it on in "Prices and discounts", it sends your description, through an ai_query statement, to the serving endpoint of your workspace that you name, and adds nothing without your click. If that endpoint serves an external model, Databricks passes the description on to its provider, and a warning tells you so. Databricks bills these calls on your account.

What about the lineage and costs Databricks already shows?

YALineo reads the same system tables and follows their definitions. It counts the direct neighbors first, as the Lineage tab of Catalog Explorer does, then the whole chain, under the label "Along the whole chain". A read through a view counts once. When a figure differs from a Databricks screen, like the "Popularity" of Catalog Explorer, YALineo says why. It then links what Databricks shows separately: the whole workspace on one map, the spend attributed to each table and the estimate of a workload before it exists.

Where do the Calculator's prices come from?

From dated snapshots of public prices bundled in the extension: Microsoft's price data for Azure, Databricks' and AWS' for AWS, Databricks' for GCP, and, for services, the pricing pages of Databricks and Microsoft Learn. The extension never downloads them; the snapshots change with the versions. On request, "Read my account's prices" reads the list prices of your account (system.billing.list_prices), which then come before the public prices. Your discounts and the prices you enter come on top, and every price says where it comes from.

What are the profiles for?

To focus on your role. At the "Your profile" step of the settings, each person picks "Full" (the default, everything open) or one of seven profiles: business, data steward or owner, data developer, data architect, FinOps lead, data scientist, BI analyst. A profile puts its pages first and a padlock on the others, which stay in the menu. The profile does not control access: anyone can change it, and YALineo always reads the data with your own Databricks permissions.

Does YALineo show what Genie costs?

Yes, from the billing. Each Genie line carries the product Databricks gives (Genie Code, Genie One or Genie Agents), and YALineo counts the free usage separately. It ties the questions of Genie Agents on SQL warehouses to each agent; they are already part of the warehouse's cost, and YALineo never adds them twice. YALineo does not use Genie. It reads the billing, the query history and the names of the Genie Agents, read only.

Can I change how it looks?

Yes. In the settings, "Appearance" offers the theme (light, dark or the one of the Databricks page) and eight styles, among them the default "YALineo" style and two Glass styles. The colors of the lines, series and states never change, and the map keeps its drawing. Where a background needs it, text and the diagnostic disc only take a shade of the same hue to stay readable.

Why an extension rather than an online service?

So that your metadata goes through no YALineo server: the extension reads your workspace from your browser. It installs from your browser's store, works with your session and needs no account.

Can we read the source code?

Yes, on request, for a security review: write to contact@yalineo.com.

What about Microsoft Fabric?

It is the planned next step. The map and the FinOps side are designed to take a second source. Fabric has its own lineage and bills its capacities in compute units. The Databricks version comes first.