Closed franksc closed 2 years ago
This only limits the total number of connections at a point in time, rather than guarding against individual clients making too many connections. Rate limiting (n requests per second/minute) should be applied at the API instance level.
Adding nginx (port 80) in front of API service (move from port 80 to 8000) has allowed for simple per-origin rate limiting
There should be a rate or connection limit for incoming API requests, to guard against: