Skip to content

Elevated Rate of DEADLINE_EXCEEDED #407

@PeterLavetsky

Description

@PeterLavetsky

For quite some time we've been making SearchStream calls with a GrpcCallContext timeout duration of 5 minutes and this has served us well.

In the last two days, over a wide range of customers and queries, we've seen a much higher rate of calls timing out with DEADLINE_EXCEEDED:

./new-google-requests-detail.2021-02-22.log : 0
./new-google-requests-detail.2021-02-23.log : 4
./new-google-requests-detail.2021-02-24.log : 2
./new-google-requests-detail.2021-02-25.log : 0
./new-google-requests-detail.2021-02-26.log : 1
./new-google-requests-detail.2021-02-27.log : 0
./new-google-requests-detail.2021-02-28.log : 28
./new-google-requests-detail.2021-03-01.log : 3
./new-google-requests-detail.2021-03-02.log : 0
./new-google-requests-detail.2021-03-03.log : 1892
./new-google-requests-detail.2021-03-04.log : 1001

We've been on com.google.api-ads:google-ads:11.0.0 for the duration of the above instance counts.

Request logs deleted

The scope does not seem to be limited to any customer or query.

No RequestIDs available as they're not returned when the Deadline Exceeded is thrown. Is it possible to make those available in a future version?

Thanks
Pete

Metadata

Metadata

Assignees

No one assigned

    Labels

    bugSomething isn't working

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions