increase read timeout for larger discovery docs #192
Conversation
| @@ -61,7 +61,7 @@ InputStream postRequest(String url, String boundary, String content) throws IOEx | |||
| HttpRequestFactory requestFactory = new NetHttpTransport().createRequestFactory(); | |||
| HttpRequest request = requestFactory.buildPostRequest(new GenericUrl(url), | |||
| ByteArrayContent.fromString("multipart/form-data; boundary=" + boundary, content)); | |||
| request.setReadTimeout(60000); // 60 seconds is the max App Engine request time | |||
| request.setReadTimeout(600000); // Larger discovery documents require more than 20 sec to process | |||
There was a problem hiding this comment.
This is surprising. More than a minute seems like something has gone wrong and the request should time out. None of our docs should be nearly large enough to go over a minute.
There was a problem hiding this comment.
I think it's just a combination of a larger discovery doc and a smaller compute core. We gradually hit the limit once in a while, with increasing frequency as our discovery doc grew. We have 295 endpoints currently.
There was a problem hiding this comment.
@elharo We've since reduced our discovery doc, but still run 50 sec and hit the limit once in a while with the upstream version. Are there any internal things you'd need to check on why this takes so long? From our perspective, increasing the timeout should be fine. Anyway, in case increasing the timeout is not viable upstream, the comment 60 seconds is the max App Engine request time is not true anymore, the max App Engine request time is now 10 min.
This fixes GoogleCloudPlatform/endpoints-framework-gradle-plugin#99
It seems, the current deployment of
google-api-client-libraries.appspot.comis not limited to the GAE Standard 60 sec request timeout anymore.