Documentation

General Questions

CDS does not return inventory data itself; it guides clients on what to scan.

Who should use the CDS API?

The CDS API is intended for:

• Live scanning clients looking to optimize efficiency and reduce load
• Cache-based clients planning to migrate to a live scanning approach
• Clients managing large inventories where full scans are costly or inefficient

Does CDS replace inventory scanning?

No.
CDS does not perform scanning and does not replace your existing scanning logic.

It complements your system by recommending where to focus your scans.

Functional Behavior

How often should I call the CDS API?

There is no fixed rule. It depends on:

• Your scanning capacity
• Inventory volatility
• Business needs

Typical patterns include:

• Maximun recomended of 1 request every 5 seconds
• Minimum depend of your capacity, recomended 1 request every 5 minutes.

Are recommendations always accurate?

CDS provides best-effort recommendations based on detected changes.
Clients should treat them as guidance, not guarantees.

What happens if I miss a recommendation?

Missing a recommendation does not break consistency, but it may:

• Delay discovery of certain changes
• Reduce scanning efficiency temporarily

CDS is designed to work as a continuous loop, not as a strict event system.

Performance & Limits

Is there a rate limit?

Yes.
Rate limits apply to protect the service and ensure fair usage.

Details (limits, headers, behavior on exceed) are defined in the API reference or communicated during onboarding.

What should I do if I receive HTTP 429 (Too Many Requests)?

• Reduce request frequency
• Implement backoff and retry logic
• Review your polling strategy

Persistent issues should be discussed with the CDS team.

Can CDS return empty recommendations?

Yes.
An empty response means no changes were detected at that moment.

This is a valid and expected scenario.

If no hotels are configured for the client, CDS will also return empty response.

Can I combine CDS with my existing scanning logic?

Yes, and this is the recommended approach.

Typical pattern:

• Use CDS recommendations as input
• Apply your own business rules, prioritization, and throttling
• Keep control over execution

Support

How do I get support if something goes wrong?

Check API errors and documentation first

• Contact your technical integration or account manager
• Provide request IDs and timestamps when reporting issues