Telemetry pipeline is how you bring measures (such as cpu_utilization_pct or customer requests) and their dimensions (such as Region or Service) into Umbrella. You set one up with a guided 5-step wizard. Once the pipeline is processing, you can build KPIs on its measures and use them across the platform.

Start from the Metrics Pipelines tab of Unit Economics and click + Add New Pipeline.

The onboarding flow at a glance

Step 1: Get Started

The first screen summarizes what the wizard will walk you through. Review it, then click Get Started.

Step 2: Upload & Configuration

2.a Unit Economic Organization

2.b Pipeline Name & Load Sample file

1. Enter a Pipeline Name - a unique name that identifies this pipeline across the platform.
2. Umbrella generates an S3 bucket where all uploaded metrics files are stored for periodic uploads.

   📘

   S3 Bucket

   Your S3 Bucket is automatically assigned by Umbrella. All uploaded metrics files are stored here. Use the copy icon if you need it for periodic uploads.

   

3. Retrieve your Access Key using the get the access key via this link link.

   ❗️

   Save your access key immediately

   The access-key link expires after one view or 5 days. Save the key somewhere secure - it's required for uploading data, and you can't view it again afterward.

   

4. Under Upload Sample File, click or drag a representative CSV. Umbrella detects the columns and generates a preview with a row count.

   ⚠️

   Samle File Prerequisite

   • Umbrella only needs it to detect columns and preview the data; you don't upload your full dataset here.
   • Files can contain up to 1,000 rows.

5. Click Next.

Step 3: Review & Mapping

Umbrella displays a File Preview (including sample rows and detected row and column counts) to help you confirm the file was parsed correctly. You can then map each column to a pipeline role: Date, Account ID, Measure, or Dimension (Cloud/Custom Dimension).

⚠️ Required Dimensions & Mappings

Four mappings are mandatory:

⚠️ Dimensions Values Requirements

The pipeline data mapped to Region, Family Type, and Services dimension values must match Umbrella's predefined values. Use the list below to align your injected data with the values used to match cost data.

Step 4: Measure Config

Define your measures - the numeric values you're tracking (for example, throughput_mbps, storage_gb, Memory_util). For each measure, set an aggregation function (Sum, Average, Min, or Max) and assign a unit. You can configure up to 5 measures per pipeline. Click Next.

Step 5: Summary

• Review the full configuration - pipeline name, file name, S3 bucket, loading method, row count, columns detected, measures, and dimensions.
• Keep the email notification option selected (and confirm the address) if you want to be told when processing finishes.

• Click Finish to validate the file and start processing.

---

After processing: success or failure email

When processing finishes, Umbrella emails the address you configured.

• On success, you'll get a confirmation that processing is complete, with shortcuts to Open Umbrella Cost or Create KPIs. You can now start monitoring metrics, creating KPIs, and scheduling the data for periodic loading.

  

• On failure, you'll get an email describing the issue with a Verify your pipeline configuration link, plus a path to support if you need help.

  

---

Load more data later

Once a pipeline exists, keep it fed via UI upload or the S3 file load.

• Update existing data by uploading the same dimensions + timestamp with new values.
• Delete data by sending the same time range with null values.

---

Limits and reference

Editing Pipelines

Click the edit icon to modify the pipeline, including its name, column mappings, measure names, aggregation, unit, and category.

Manual Data injection

Upload data via the UI by clicking the icon and selecting a file.

Pipeline Data Validation

• The system checks your file content against the pipeline configuration, column names, datetime format and account id values.
• Missing mapped columns will fail the process; extra unmapped columns are safely ignored.

Processing

Once validated, metrics data processing begins. You'll receive an email notification on success or failure.

MSP scenarios

• MSP users can inject Unit Economics telemetry on the MSP organization.
• MSP users can inject telemetry into a customer organization on the customer's behalf.
• MSP customers can inject telemetry into their own customer organization.

Cloud dimensions