#Quotas and Limitations

To ensure optimal performance and system stability, PX Insights enforces certain technical limitations. Understanding these limits will help you design an effective integration strategy.

#Queue-Based Processing

PX Insights uses an asynchronous queue-based system to process review data. This approach offers significant advantages:

• High throughput: You can send large volumes of reviews without worrying about overwhelming the system.
• Reliable processing: The queue ensures all reviews are processed, even during peak loads.

While there are technical limits outlined below to protect system stability, the queue-based architecture allows for high-volume processing.

#Payload Limitations

To maintain system performance, we enforce the following constraints on request payloads:

Parameter	Limitation	Notes
Maximum requests per second	1000 requests	Per client ID
Maximum reviews per request	500 reviews	For larger batches, split into multiple requests
Supported score range	1-5	Integer values only (1, 2, 3, 4, 5)

#Processing Time Expectations

While PX Insights processes reviews asynchronously, you can expect the following processing times under normal conditions:

• Standard processing: Reviews are typically processed within 5-10 minutes.
• High-volume periods: During peak times, processing may take up to 30 minutes.
• First-time imports: Initial data loads or large migrations may take longer, depending on the volume of reviews to perform ID reconciliation.

#Troubleshooting Missing Reviews

If your reviews don't appear in the PIM after the expected processing time (30+ minutes), it's likely due to an ID reconciliation failure:

#Common causes:

1. Missing product identifier: Ensure you've included at least one product identifier (preferably sku) in the metadata object.
2. Invalid identifiers: Verify that the product identifiers in your requests exactly match those in your PIM.
3. Product not found: Confirm that the referenced products actually exist in your PIM catalog.

#Resolution steps:

1. Check your request payload to ensure the product_identification section includes valid metadata.
2. Verify that your origin field is correctly specified.
3. For direct mapping, ensure both product_type and either product_uuid or product_model_code are accurate.
4. For metadata-based matching, confirm that your metadata contains identifiers that can be found in your PIM.

#Best Practices

To optimize your experience with PX Insights:

1. Filter by relevance: Consider filtering your reviews by date. Reviews lose relevancy over time, so you might not need to import everything.
2. Avoid duplication: Only import new data rather than re-importing the same reviews. Our API handles duplicates well, but unnecessary re-imports consume resources.
3. Batch sensibly: While you can send up to 500 reviews per request, consider using smaller batches (20-50) for more predictable processing.
4. Review processing status: For large imports, consider implementing a monitoring mechanism to track overall progress.