Query/response limit

BLUF: Is the AeroAPI definition of “query” just API calls, or API calls + response pages/sets?

I’m currently using the personal tier of AeroAPI, which limits users to 10 queries/minute. I’m requesting SkyWest departures from Denver using “/airports/KDEN/flights/departures”. If I set max_pages to more than 8, I get a “429 - Too many requests” response. Are the number of records/response pages I can receive from an API call capped by the 10 queries/minute limit? I’m not making more than 1 API call per minute, which leads me to believe that the response pages/records returned by the API count towards the query limit, however I can’t find any documentation that covers this. Thanks in advance for any help!

Looking at how the pricing structure is for AeroAPI v4 will provide a clearer look as to how the “pages” are broken down, but to answer your question yes a query would be a single page for calls that can request multiple pages.

Taylor Fischer -
Commercial Support

Toll Free: +1.800.713.8570
Support Line:+1.713.588.6514

Eleven Greenway Plaza, Suite 2900 - Houston, Texas - 77046
FlightAware


Learn about FlightAware Global and Premium Accounts: Commercial Services - FlightAware

Thanks for the quick response. Just so I understand, if my API call returns 5 result sets, that counts as 6 queries (in terms of query rate limit)?

I am using the Personal Tier and having the same issue when requesting departures and arrivals for specific airports. Has the your theory of original call + response page/record count = query count been confirmed?

On a side note, based on my API Usage graph, a user is charged for the “429 - Too many requests” response. Just curious, do the original call and the 429 response add to the query count as well?

@bauergda @bneckfootball38

Rate limiting is generally performed at the result set level. Therefore, a query that would return more than 10 result sets is not possible on the Personal tier. To answer the specific question from @bneckfootball38, a call returning 5 result sets consumes 5 quota. A call with 0 results still consumes 1 quota, as does any 4xx error response besides a 429.

Some more examples:
0 results → 1 total quota consumed
15 (individual) results → 1 quota
16 results → 2 quota
404 error → 1 quota
400 error → 1 quota
429 error → 0 quota

To answer the question from @bauergda about billing: we’ve recently adjusted our billing approach to not bill for any 4xx errors (besides 404s). This is currently not properly reflected on the usage table which we intend to address soon.