API Rate Limit
This tutorial introduces developers to Bazaarvoice API rate limiting, explains how to determine the rate limit and offers several suggestions on how to avoid exceeding the rate limit.
Overview
Usage of the Conversations API is rate limited on a per API key, per minute basis. Each key has its own limit and using one key does not impact the limit of a different key. Initially keys are set to a default limit, which should be
sufficient for most of our clients and we will pro-actively raise it during the holiday to accommodate the traffic increase across our network.
Why rate limit
The Bazaarvoice platform is a multi-tenant system in which many clients share the same system resources. As a result it is important that excessive usage is limited to protect the stability of the network for everybody. Without a rate limit, poorly constructed applications or unforeseen circumstances could have negative consequences on network performance and availability.
Determining rate limit
There are two ways to determine rate limit:
Bazaarvoice Portal
Only keys requested using our API key request process will be accessible in the Bazaarvoice Portal.
Refer to API key management to learn how to view details about API keys in the Bazaarvoice
Portal.
HTTP headers
The Conversations API will return the following headers, which your application can use to ensure that it does not exceed the rate limit.
Name | Description |
---|---|
X-Bazaarvoice-QPM-Allotted | The maximum QPM (queries per minute) capacity set on your API key. This is the maximum number of calls that can be made in any given minute. |
X-Bazaarvoice-QPM-Current | The current count of calls being applied against the above limit; in this case, it is a representation of how many calls your key is making at that particular minute in time. |
X-Bazaarvoice-QPS-Allotted | ⚠️ QPS has been superseded by QPM. New applications should ignore QPS and existing apps should stop relying on it. The maximum QPS (queries per second) capacity set on your API key. This is the maximum number of calls that can be made in any given second. |
X-Bazaarvoice-QPS-Current | ⚠️ QPS has been superseded by QPM. New applications should ignore QPS and existing apps should stop relying on it. The current count of calls being applied against the above limit; in this case, it is a representation of how many calls your key is making at that particular second in time. |
X-Bazaarvoice-Quota-Allotted | The maximum number of calls that can be made on an hourly basis. |
X-Bazaarvoice-Quota-Current | The current count of calls made in the current limit period. |
X-Bazaarvoice-Error-Detail | An error message returned when the QPM or quota is exceeded. |
X-Bazaarvoice-Quota-Reset | The time when the rate limit will reset to 0. |
In the following example the rate limit has been exceeded (X-Bazaarvoice-QPM-Current
is greater than X-Bazaarvoice-QPM-Allotted
).
HTTP/1.1 200 OK "Content-Type":"application/json", "Date":"Mon, 24 Apr 2017 20:02:46 GMT", "Server":"Apigee Router", "X-Bazaarvoice-error-detail":"Account Over Queries Per Second Limit", "X-Bazaarvoice-QPM-Allotted":"60", "X-Bazaarvoice-QPM-Current":"60", "X-Bazaarvoice-QPS-Allotted":"1", "X-Bazaarvoice-QPS-Current":"1", "X-Bazaarvoice-Quota-Allotted":"7200000", "X-Bazaarvoice-Quota-Current":"63", "X-Bazaarvoice-Quota-Reset":"2017-04-24T21:00:00.000Z", "Content-Length":"250", "Connection":"Close"
Do not rely on the order or capitalisation of headers.
How to avoid exceeding the limit
Here are some suggestions on how to avoid exceeding the rate limit.
- Use API keys appropriately
API keys should be used on a per application basis. For example you may have an API key for an application that shows user generated content (UGC) on a social network and a different API key for your eCommerce site. It is also
appropriate to use one API key per locale or region. A client that serves traffic to users in the United States and Europe might have two API keys, one for each region.
- Make requests efficiently
Making unnecessary requests is a common reason applications exceed the rate limit. A typical product page should only need one request to load the first page of UGC and then only make subsequent requests as users indicate their
desire to see alternate views (e.g. pagination, filtering, sorting, etc). Further, secondary content like review comments and answers to questions can be requested along with the primary content using the &include
parameter.
- Local caching
If you're already using your API keys appropriately and making efficient requests you might try caching frequently needed resources. The exact implementation is up to each client, but could involve pre-rendering product pages with UGC included, saving UGC in a data store or the users browser as it is requested, or hosting a caching proxy between your application and the Bazaarvoice platform. Refer to our caching tutorial to learn more.
- Store data you need
Serve UGC to your end-users from your data store, which is created and updated using data from the Bazaarvoice platform independent of end-user demand. This gives you complete control of the requests you make to Bazaarvoice and how you use your data.
Approval required
This option is limited to clients with prior approval. Contact your Bazaarvoice representative or our Support team if you would like to use the Conversations API to create your own UGC data store.
- Increase the limit
Finally you may contact your Bazaarvoice representative or our Support Team if you feel that the limit is
insufficient for your needs. Rate limit increases must be approved and may
require additional charges.
Updated 9 months ago