Marketo API errors 1003 and 603: Checks and recovery
When a Marketo integration stops, check both the error code and the request conditions. Changes in 2026 affect activity extraction from large static lists and how access tokens are supplied. Verify data completeness before interpreting a decline as marketing performance.
This guide helps marketing teams prepare a repair request and restore reporting.
Error 1003: Check list-based extraction
Since September 30, 2026, Get Lead Activities and Get Lead Changes requests using listId fail with error 1003 when the target static list contains 10,000 or more leads. This does not prohibit all lists or API usage. Adobe migration guide
Consider smaller extraction lists, or Bulk Activity Extract and Data Streams joined to list membership. Before changing the integration, check granularity, available history, usage limits, and storage with its owner.
The guide's splitting recommendation uses inconsistent boundary wording. Do not assume a list of exactly 10,000 will work. Verify responses and discuss keeping extraction lists below the failing threshold.
Error 1003 also describes other invalid data. Match the endpoint, message, and parameters before identifying this list restriction as the cause. Error code reference
Error 603: Separate authentication transport from permissions
Support for passing access_token in a REST request's URL query ended on August 31, 2026. Adobe directs integrations to the Authorization header and notes that old workflows may encounter 603. Authentication guidance
Error 603 can also indicate access restrictions. Error 601 indicates an invalid token; 602 indicates an expired token. Treating every 603 as expiration can leave the actual cause unresolved.
| Symptom | Information for the integration owner |
|---|---|
| 1003 | Endpoint, message, use of listId, list membership count |
| 601 or 602 | Token generation, refresh, and expiration handling |
| 603 | Token transport, API user permissions, access restrictions |
| Only some jobs fail | Per-job settings and last successful completion |
The removal concerns token transport in REST calls. Do not assume the same change applies to token issuance.
HTTP success does not establish successful extraction
Marketo response-level and record-level errors can return HTTP 200. An integration checking only HTTP status can overlook failure. Response and record errors
Use this internal recording worksheet; it is not a reproduction of Marketo's interface.
| Record | Purpose |
|---|---|
| Job and owner | Identify who can repair it |
| Last fully extracted interval | Identify usable reporting coverage |
| HTTP and response-body result | Separate transport from processing |
| Extracted and stored counts | Verify completion through storage |
| Missing intervals | Plan recovery and annotate reporting |
| Deduplication method | Prevent duplicate outcomes after replay |
A low count with confirmed successful extraction can be investigated as a business change. A failed extraction interval should first be marked as missing.
Restore downstream reporting as well
- Identify affected jobs, reports, and audience conditions.
- Record the last successful and first failed intervals.
- After repair, verify extraction through storage within a limited scope.
- Agree on recoverable history and deduplication before backfilling.
- Verify updates to summaries, AI analysis, and audience rules.
Splitting extraction lists should not unintentionally change delivery or sales audiences. Replaying extraction and replaying email delivery are separate operations.
Include data status in AI analysis
Analyze this marketing summary.
Source and reporting period:
Successfully extracted coverage:
Missing, partially failed, or recovering intervals:
Metrics and business conditions:
Do not treat missing data as zero outcomes.
Separate observed changes from conclusions blocked by missing data.
Separate data extraction checks from marketing performance checks.
Give the integration owner the error and request conditions, then annotate reporting coverage. Keep that annotation when sharing marketing data with AI. Shoryaku AI can discuss summaries and business information you supply; it does not connect to Marketo or repair its APIs.