We’re just seeing a fair amount of these during certain times and cannot track down what it is.
a) client makes call to simple service via Tyk. This backend service who when called directly and not via tyk responds consistently in < 100ms
b) the plugin calls TykGetKeyData (to lookup a session), if session is valid (in this case it is) it proceeds to proxy the request.
c) all of the log statements generated by the plugins for this request all share “Jul 19 09:09:26” (1 second). So I assume the plugins are executing in < 1 second
d) yet the last log statement (from tyk) shows the client timing out… .(the client has a 30s timeout) “http: proxy error: context canceled”. The timestamp on this tyk log statement is the same “Jul 19 09:09:26”
e) again the backend service is responding fine.
A bit lost as to what is causing this, its definitely something in tyk but not sure where to look as the logs are not very helpful
I can reproduce this in a manual call, its like one in every 30 requests has this issue. The backend responds fine, but when I go through tyk this happens every so often