Skip to main content

Upload and Manage Developer Records Using the Harness SEI API

Overview

This page explains how to create and update developer records programmatically, provides best practices for integrating with external systems, and includes troubleshooting information to validate developer data before applying changes.

For a complete list of endpoints and schema definitions, see the Harness API reference documentation.

Before you begin

To call the SEI API, you’ll need:

  • A Harness API key

  • A Harness account identifier

  • A base URL. All API URIs are relative to the SEI service endpoint for your region.

    • Prod 1: https://app.harness.io/prod1/sei/api/
    • Prod 2: https://app.harness.io/gratis/sei/api/
    • EU: https://accounts.eu.harness.io/sei/api/

Authentication

All Harness SEI APIs require authentication using a Harness API key.

To authenticate:

  1. Create a Harness API key.

  2. Include the API key in your request header:

    curl -X GET \
    "https://app.harness.io/gateway/sei/api/v2/developers/schema" \
    -H "x-api-key: <YOUR_API_KEY>" \

Manage developer records

The Harness SEI API enables programmatic management of developer records. You can use the API to automate updates by writing scripts or jobs that trigger record changes whenever your source of truth (such as an HR system) changes.

This reduces manual edits, lowers the risk of errors, and ensures developer data stays accurate and consistent for productivity and DORA metrics.

Initial setup (Upload and save)

The developer schema and baseline data must first be established using the upload and save workflow:

  1. POST /v2/developers/upload: Upload a CSV file containing developer data (CSV only, cannot be JSON), which returns an upload preview with field mappings.
  2. POST /v2/developers/uploads/{uploadId}/save: Configure field mappings (such as NAME, EMAIL, MANAGER_EMAIL) and save developers into the system. You must complete this step before using other developer management APIs.
danger

Reupload, approve, and upsert operations will return 412 Precondition Failed if attempted before completing this initial setup.

Bulk update (Re-upload & approve)

To update existing developer data with a new CSV file, the initial upload/save must be completed first.

  1. POST /v2/developers/reupload: Upload a new CSV file with updated developer data. This automatically uses existing field mappings and returns review data showing changes.
  2. POST /v2/developers/uploads/{uploadId}/approve: Review and approve the changes, and activate the new developer dataset.

Developer endpoints

ActionEndpointUsage
Reupload developer recordsPOST /v2/developers/reuploadReupload a CSV file with review data for approval workflow. CSV file only.
Approve developer recordsPOST /v2/developers/uploads/{uploadId}/approveApprove developers reupload and activate the changes.
Upsert developersPATCH /v2/developersAdd new developers or update existing ones. JSON only, no two-step approval process.
Delete developersDELETE /v2/developersDelete developers by their email addresses.
Search / list developersPOST /v2/developers/listSearch developers with filtering and sorting options. Search by attributes (key-value pair like "department":"Engineering", name).
Get developer schemaGET /v2/developers/schemaGet the latest developer field mappings (schema) for the account. Retrieve the schema for developer records and field mappings.
Download developer dataGET /v2/developers/downloadDownload a CSV file containing all developers currently in the system for auditing or sync back to third-party systems.

cURL commands

Reupload developer records (POST /v2/developers/reupload)

curl -X 'POST' \
'https://app.harness.io/prod1/sei/api/v2/developers/reupload?fileType=CSV' \
-H "accept: application/json" \
-H "Content-Type: multipart/form-data" \
-F "file=@Sample_CSV_for_Importing_Developers.csv;type=text/csv"

Approve developer records (POST /v2/developers/uploads/{uploadId}/approve)

curl -X 'POST' \
'https://app.harness.io/prod1/sei/api/v2/developers/uploads/<YOUR_UPLOAD_ID>/approve' \
-H 'accept: */*' \
-d ''

Upsert developers (PATCH /v2/developers)

curl -X 'PATCH' \
'https://app.harness.io/prod1/sei/api/v2/developers' \
-H 'accept: */*' \
-H 'Content-Type: application/json' \
-d '[
{
"<NEW_RECORD>": "string",
"<NEW_RECORD>": "string",
"<NEW_RECORD>": "string"
}

Delete developers (DELETE /v2/developers)

curl -X 'DELETE' \
'https://app.harness.io/prod1/sei/api/v2/developers' \
-H 'accept: */*' \
-H 'Content-Type: application/json' \
-d '{
"emails": [
"john.doe@example.com",
"jane.smith@example.com"
]

Search/list developers (POST /v2/developers/list)

curl -X 'POST' \
'https://app.harness.io/prod1/sei/api/v2/developers/list
pageIndex=<INTEGER>&pageSize=<INTEGER>' \
-H 'accept: application/json' \
-H 'Content-Type: application/json' \
-d '{
"sortOrder": "asc",
"searchKey": "string",
"searchValue": "string",
"sortKey": "string",
"developerRefIds": [
1073741824
]

Get developer schema (GET /v2/developers/schema)

curl -X 'GET' \
'https://app.harness.io/prod1/sei/api/v2/developers/schema' \
-H 'accept: application/json' \

Download developer data (GET /v2/developers/download)

curl -X 'GET' \
'https://app.harness.io/prod1/sei/api/v2/developers/download' \
-H 'accept: text/csv' \

Sync developer records

You have two options for syncing developer records with Harness SEI APIs:

  1. Event-Driven Sync (JSON, small updates): This is best for event-driven syncs, such as when a new hire webhook triggers an update.

    danger

    Upsert will return 412 Precondition Failed if the initial setup hasn’t been completed.

    • Fetch developer records from an external source such as an HR system.
    • Call PATCH /v2/developers to insert or update the records.
    • Validate using POST /v2/developers/list.
  2. Scheduled Sync (CSV, bulk updates): This is best for nightly or weekly HR exports. CSV is required for upload and save, and JSON is not supported.

    • Upload a CSV file using POST /v2/developers/reupload and POST /v2/developers/uploads/{uploadId}/approve (response includes a preview and validation).
    • Review preview or validation errors. Correct upstream as needed.
    • Save developers using POST /v2/developers/uploads/{uploadId}/save to commit to the system.
    • Export the latest developer list using GET /v2/developers/download for auditing.

For automated syncs, Harness recommends running them daily in high-churn orgs and weekly in stable ones, with on-demand syncs when HR system changes occur. Always preview and audit changes, validate errors early, and apply business rules for invalid rows.

Example use case

A nightly job queries an HRIS system for all active employees in engineering.

  1. Upload a CSV file containing developer data using POST /v2/developers/upload. This returns a preview with field mappings and validation errors.
  2. Check for validation errors in the preview (e.g. REQUIRED_FIELD_MISSING or DUPLICATE_EMAIL) and correct errors upstream in the CSV file if necessary.
  3. Commit the uploaded developer data into the system using POST /v2/developers/uploads/{uploadId}/save. The initial setup must be completed before using other APIs, otherwise upsert, reupload, or approve operations will return 412 Precondition Failed.
  4. Optionally, export the current developer list using GET /v2/developers/download for auditing or syncing with external systems.

Troubleshooting

Developer operations (save, upsert, delete, and reupload) perform comprehensive validation and return structured error responses when validation fails.

Validation errors help identify and resolve issues with CSV uploads, field mappings, and developer record data before any changes are applied.

Error types

Error TypeWhen it OccursDetails
REQUIRED_FIELD_MISSINGA required field is missing from the CSV upload or field mappings.Required fields include name, email, and manager_email (if configured), as well as any fields used in org tree filters or group-by configurations. Triggered during file upload when mappings don’t include required fields.
REQUIRED_VALUE_MISSINGRequired fields have empty or blank values in developer records.Applies to name and email (always required). Triggered when any developer record has empty/null values for these fields.
DUPLICATE_EMAILMultiple developer records share the same email address.Each developer must have a unique email. Applies to all developer operations (save, upsert, delete).
CIRCULAR_REFERENCEManager-employee relationships form circular references.Example: A reports to B, and B reports to A. Detected through graph traversal of manager-employee relationships. Prevents infinite loops in the org hierarchy.
MANAGER_EMAIL_MISSING_IN_DEVELOPER_EMAILSA developer’s manager_email value doesn’t exist in the dataset.Manager email can be empty, but if provided, it must match an existing developer’s email. Ensures all managers are also developers.
MISSING_NAME_OR_EMAIL_OR_MANAGER_EMAIL_FIELD_MAPPINGEssential field mappings (NAME, EMAIL, or MANAGER_EMAIL) are missing from field configuration.Triggered during save when required field types are not mapped to CSV columns. Prevents saving when core identification fields are undefined.
INVALID_CSVThe uploaded CSV file format is invalid or corrupted.Triggered during upload when parsing fails. Causes include malformed structure, encoding issues, or file corruption.
EMPTY_CSVThe uploaded CSV file contains no data rows.Triggered when the file has only headers or is entirely empty. Prevents processing of empty datasets.
BLANK_CSV_COLUMN_NAMEThe CSV contains columns with blank or empty header names.Triggered during upload when headers are missing or contain only whitespace. Ensures all columns have valid names for mapping.

Error response structure

All validation errors are returned in the following structure:

{
"status": "FAILURE",
"errors": [
{
"errorType": "ERROR_TYPE_NAME",
"fieldName": "field_name_or_null",
"errorDeveloperCount": 10,
"sampleErrorDevelopers": [
{
"field1": "value1",
"field2": "value2"
}
]
}
]
}
  • sampleErrorDevelopers is limited to a partial list of records per error type.
  • errorDeveloperCount shows the total number of records affected.

Multiple error types can be returned simultaneously and validation occurs before any data is saved to prevent partial updates.