the Developer creates an account and request an access to the API
the Admin or a method validate the access
the Developer receives a token to use in his app
the Developer use the tokens to setup its own client application consuming the API
Before releasing this, we should be able to monitor the following data :
Number of API calls
API response times by route
API and workers load
We should be able to filter this data for each token / app / client in order to prevent abuse of the infrastructure and eventually scale.
We should also add django methods or API calls or UI to let admin remove a client access
We should also think about a ratio system (by API calls ? by worker's processing time ? By number of analyzed seconds in an audio file ?)
As talked in #165 we should define (and implements in the future) the workflow to provide API tokens to the users of the Timeside API.
Here's a first draft adapted from @yomguy's comments
Before releasing this, we should be able to monitor the following data :
We should be able to filter this data for each token / app / client in order to prevent abuse of the infrastructure and eventually scale.
We should also add django methods or API calls or UI to let admin remove a client access We should also think about a ratio system (by API calls ? by worker's processing time ? By number of analyzed seconds in an audio file ?)
Any thoughts ?