Retry-After header lost/changed

[csimpi]

Issue found on February 19th 2015.

Endpoint(s):

• Each

Scope(s):

• Not relevant

Steps to reproduce:

1. Trigger the throttle limitation somehow
2. Check the headers

Expected behaviour:

Previously I could parse the Retry-After's value from the header and wait.

Actual behaviour:

Now Retry-After is not in the headers anymore but I can get retry-after instead, but its value is always 1, after waiting I'm still getting 429 error. This has been working before for long months with no issues. Is this documented somewhere?

[csimpi]

No support here anymore?

[ghost]

@csimpi thanks for the report and sorry for the slow response. We will look into this and get back as soon as we know more.

[ghost]

@csimpi we have investigated this some more, and it looks like the retry header is set correctly in general. Could you please provide more information for your specific case?

[travis-white-6]

I am not seeing the Retry-After header at all but am getting a response 429 every once and a while - strangely enough it's always at the start of the hour and 11 seconds in. Two recent examples are:

response.status_code=429 with response.headers={'Server': 'nginx', 'Date': 'Mon, 23 Sep 2019 12:00:11 GMT', 'Content-Length': '0', 'Connection': 'keep-alive', 'Keep-Alive': 'timeout=600', 'Set-Cookie': '__Host-device_id=AQCw3Rkx8I3pAJdo3yCHdjrcGhLLuDSASxkyX-EDT6onM4W5siW5JwVN1HIhwhDnvkYWbNd1eu9jSOoO6cMop7pdf0B_YfuoVmU;Version=1;Path=/;Max-Age=2147483647;Secure;HttpOnly'}

response.status_code=429 with response.headers={'Server': 'nginx', 'Date': 'Sun, 22 Sep 2019 13:00:11 GMT', 'Content-Length': '0', 'Connection': 'keep-alive', 'Keep-Alive': 'timeout=600', 'Set-Cookie': '__Host-device_id=AQDoR5bTmtjUrxyYML2bd4eW40Rws9b1KxD91HmQtuPNL7Oy7kuF_1ssxUX3-qYginZSL4DQToVp3fB0YiRmRXMFKFbccvTWQlw;Version=1;Path=/;Max-Age=2147483647;Secure;HttpOnly'}

Language: Python
Using the Web API

Any suggestions / help you can offer? Right now we do not have an implementation to throttle API requests internally since we can't seem to get the Retry-After header. If we can get a value for that I'd be more than happy to self-impose a wait time in our API calls