Cisco Prime Infrastructure API
Prime Infrastructure API Documentation

Frequently Asked Questions

Additional answers may be found in the getting started section about the Prime Infrastructure API, and in the tutorial.

  1. What is the syntax for requests to the Prime Infrastructure API?
  2. XML or JSON in the response?
  3. How do I reduce API latency and increase throughput?
  4. Why am I getting HTTP 401 errors after 5 successful connections?
  5. Why am I sometimes receiving HTTP 503 errors?

1. What is the syntax for requests to the Prime Infrastructure API?

The general structure of a request to the Prime Infrastructure API is as follows:

  ../webacs/api/v1/data/{RESOURCE-NAME}
            

This will list all instances of that resource. For example, ../webacs/api/v1/data/Devices will return a list of all Device instances in the system.

  ../webacs/api/v1/data/{RESOURCE-NAME}/{RESOURCE-ID}
            

This will return the unique instance identified by the given {RESOURCE-ID}. For example, ../webacs/api/v1/data/Devices/12412764 returns the "12412764" device.

Qualifiers may be added to use filtering, paging and sorting following this format:

  ../webacs/api/v1/data/{RESOURCE-NAME}?{operator-expression1}&{operator-expression2}...&{operator-expressionN}
            

Each operator expression may be a filtering criteria such as ipAddress="192.168.1.3", or a paging directive, or a sorting directive, such as .sort=-collectionTime.

Optionally, the .full=true qualifier may be added to return the full instance.

Additional RESTful services are available through the following URL pattern:

  ../webacs/api/op/{service-domain}/{service-resource}            
            

For example, ../webacs/api/op/reportservice/templates provides a way to list all report templates that may be used to create reports.

2. XML or JSON in the response?

The response to a request to the Prime Infrastructure API may be returned in one of two standard formats: XML or JSON. The format of the response may be specified in one of two ways:

  • By adding ".json" or ".xml" to the URL, at the end of the path but before the query string. For example:
      ../webacs/api/v1/data/Devices.json?ipAddress="192.168.8.9"              
                  
    returns a response in JSON format, whereas
      ../webacs/api/v1/data/Devices.xml?ipAddress="192.168.8.9"              
                  
    returns a response in XML format.
  • By specifying the expected format in the request "Accept" header:
      Accept: application/json
                  
    returns a response in JSON format, whereas
      Accept: text/xml
                  
    returns a response in XML format.

By default, when nothing is specified, the response format is set to be XML.

3. How do I reduce API latency and increase throughput?

If you would like to cut down the latency of API requests, or improve throughput, you can try some of the following tips:

  • For queries against resources that support paging, try using the ".nocount" parameter and set it to true. For example
      ../webacs/api/v1/data/Devices?.nocount=true
                 
    Doing so will remove the "count," "first," and "last" attributes from the API response. However, the API will require one less query against the underlying data source, yielding a lower latency.
  • Consider using Summary resources if they report all the information you need. API latency can be reduced by using resources that return more limited sets of information. For instance, if you wish to get the IP addresses of clients, then Client Summary will report the data you need faster than Client Details.
  • Use larger page sizes for resources that support paging. While it might seem counter-intuitive, most of the performance costs paid by the API are on a per-request basis, not a per-record basis. By increasing page sizes, you can reduce the overall number of requests and increase your overall throughput.

4. Why am I getting HTTP 401 errors after 5 successful connections?

You're likely sending the "Connection: Keep-Alive" header. The API doesn't presently support persistent connections. So each time the header is sent a session is created for you. However, instead of cleaning up the session after a response is sent the session sticks around. In a recent version of Prime Infrastructure, we introduced a limit of 5 concurrent sessions per user. So these two things combined means that when your sixth request comes around it is denied with an HTTP 401. The workaround is to not send the "Connection: Keep-Alive" header or disable persistent connections in your client (for some clients, sending the "Connection: close" header in your request disables persistent connections).

5. Why am I sometimes receiving HTTP 503 errors?

If you sometimes receive HTTP 503 errors, but at other times receive a success response, you are being rate limited. The best thing to do in this case is to wait for a few seconds and try the same request again. For more details, please see our rate limiting documentation.