Process terminated because the request deadline was exceeded. (Error code 123)

[Леонтьев Пётр]

Hi.

We have a lot of errors like:
InternalError: (google.storage.speckle.python.api.rdbms.InternalError) (0, u'Instance(tinyarmypanoramic:ratings) has too many concurrent requests: 100')

InternalError: (google.storage.speckle.python.api.rdbms.InternalError) (0, u'Connection is already in use.')

In the attachment you can see number of Google Cloud SQL connections during 16 June (starting from 10:00 UTC+0).

Problem looks very similar to this: https://groups.google.com/forum/#!topic/google-appengine/-yrM5IXHoaQ

[Леонтьев Пётр]

In addition, we use 1st generation Cloud SQL.

четверг, 16 июня 2016 г., 15:06:38 UTC+3 пользователь Леонтьев Пётр написал:

[Adam (Cloud Platform Support)]

This isn't related to an internal issue as per the thread you linked to - your Cloud SQL instance literally has too many concurrent requests as the message states, and there is a limit of 100. 

For some background, take a look at this answer from a Cloud SQL engineer about what to do about these kinds of errors.

[Adam (Cloud Platform Support)]

Some more info is available in the docs - "If more than 100 clients try to connect simultaneously to a First Generation instance, some of them will fail to connect". Another possible solution may be to use a higher tier instance so that queries will complete faster and reduce the concurrent load on the instance.

[Леонтьев Пётр]

Thank you for response! I'll take provided info into account.

четверг, 16 июня 2016 г., 15:06:38 UTC+3 пользователь Леонтьев Пётр написал:

Hi.