Food Hygiene Rating API

If you require regular updates of entire local authority (LA) data, consider downloading the Open Data files and then processing the data from your local copy.

These files are updated nightly. The data source is the same as the ratings website, and the V1 & V2 APIs.

Open Data is the best option to get full daily updates. The files have an extractDate in them that signal when the data has changed (not all LAs change every night). By using these files and the extract date, you can minimise your processing, and get a full update as fast as possible.

You can get the list of the URLs for Open Data files by querying the local authority endpoint in the API. It is advised that the URLs are queried from this list: there is the possibility that the files will be moved (e.g. to a Content Delivery Network) in the future, at which point the URLs will change: if you are getting them from the API the change will be seamless.

If you can't process the Open Data files and must use the APIs directly, it is a good idea to implement flow control. This helps avoid your requests being rejected due to throttling.

Consider using a sleep function to control the rate at which you send requests. It is advised to leave the rate configurable, as throttling limits may change depending on the current operational climate.

API requests with large page sizes (greater than 200) result in increasingly high resource demand: multiple concurrent large requests can have an impact on the availability of the FHRS service.

Page sizes above 200 are subject to rate limiting. If a large page request results in an Http 429 error, this is the likely cause.

If you are retrieving bulk data from the API (rather than using Open Data files), rather than attempt to retrieve all data in a single request, it is advisable to implement paging, using a page size of 200 or less, and then use mulitple requests, with flow control.

If you can run your updates outside of evening meal/takeaway times, this helps because the system isn't as loaded. Running it as a cron job after midnight is a good approach. Note that the same throttling level applies throughout the day, so this won't help if you are seeing 403/429 errors.