Sanitized Ramp Support Bundle
Overview
Create an evidence package that Ramp support and internal responders can use without exposing credentials, card details, receipts, or unnecessary employee and financial data.
Prerequisites
- Identify the Ramp application, environment, business entities, affected data and workflows, accountable owner, and rollback boundary.
- Read
references/official-docs.mdand re-check endpoint schemas, scopes, limits, and support status before a live operation. - Use synthetic fixtures or Ramp sandbox until production access and business effects are explicitly approved.
- Prepare approved secret storage and a sanitized evidence location.
Current Contract
- Ramp responses include
x-trace-id, the preferred correlation handle for developer support. - API error bodies may include
error_v2.error_code, additional information, and a message. - Webhook and deferred-task incidents need delivery/task identifiers and lifecycle state, not copied secrets or full payloads.
- Authorization failures require environment, grant, and scope metadata, but never the bearer token or client secret.
Instructions
-
Open a restricted case directory and record incident window, environment, affected endpoint or event, expected behavior, owner, and data-handling classification.
-
Capture UTC timestamps, method, sanitized path template, status,
x-trace-id, error code, attempt count, latency, idempotency key hash, and application release. -
Add a minimized schema-valid reproducer using placeholders and sandbox IDs; record grant type and scope names without credential values.
-
For webhooks or deferred tasks, include subscription/event/task identifiers, deduplication outcome, state transitions, and sanitized response metadata.
-
Run a secret/card/identity scan, obtain the data owner's sharing approval, generate checksums, and transmit only through the approved support channel.
Tool Discipline
- Use Glob to locate candidate code, manifests, fixtures, and evidence without widening scope.
- Use Grep to find relevant endpoints, fields, permissions, identifiers, errors, and stale assumptions.
- Use Read to inspect the smallest required local files and authoritative evidence.
- Use Write only for a new approved local draft, test, configuration, or evidence artifact.
- Use Edit only for a bounded approved change with a known rollback.
- Local file tools do not authorize a Ramp operation or replace owner approval.
Approval Boundaries
The incident owner approves scope; security/data owners approve external sharing. Never upload raw production bodies, receipts, PAN/CVV, tokens, or client secrets.
Output
A checksummed sanitized bundle, field manifest, reproduction steps, trace IDs, environment and release metadata, redaction scan, sharing approval, and retention deadline.
Error Handling
| Condition | Response |
|---|---|
| No trace ID was retained | Reproduce the smallest safe request in sandbox or an approved read-only path and capture the new trace. |
| Redaction destroys reproducibility | Replace values with type-preserving placeholders and keep restricted originals outside the shared bundle. |
| A support request asks for secrets | Refuse and use an approved credential-verification or secure escalation process. |
Examples
Example 1
Bundle a recurring 422 with one sanitized request fixture, error_v2 code, trace ID, schema version, and exact release.
Example 2
Document a webhook gap using event IDs, receive timestamps, dedupe decisions, and a reconciled object count without event bodies.
Validation
- The bundle reproduces or precisely bounds the failure.
- Trace IDs and identifiers are preserved while secret and sensitive value scans are clean.
- Every included field has an explicit diagnostic purpose.
- Sharing channel, approver, checksum, and deletion date are recorded.
Resources
- Official documentation and contract notes
- Re-check the dated contract and current OpenAPI schema before any live request.
- Treat unresolved vendor behavior, authority, or financial state as a stop condition.