Blobstore outage?

[rasmus]

We have seen many blobstore errors today:

DeadlineExceededError: The API call file.Create() took too long to respond

for 2 periods: around 18:15 - 18:30 GMT

and briefly around 22:00 GMT.

I have been switching between 2 versions to "fix" this: 

(1) stores 1-3 blobs for *every* request, including many <500 byte ones, and

(2) stores only >1kb blobs, on maybe 1 in 5+ requests

Version (1) had periods when most blobstore requests failed (then we respond 500, then our customers retry fairly eagerly, so it piles up)

but it also worked for a few hours,

Version (2) seems nearly problem-free - but it was of course running at different times from (1), and there were occasional rejects.

This made me wonder:

Is it possible for our app to hammer down Blobstore by storing many small blobs?

Our volumes are not huge, average 10-100 requests / minute, with peaks of 100+ requests near-simultaneously.

But in each case the outage resolved immediately when I switched from version (1) to v (2)

Thanks for any tips/confirmation.

    rasmus

[Amy Unruh]

hi,

You should not be seeing this any longer-- please post a follow-up if you do. 

(This was an issue with an experimental API -- so just as a general note, it's advisable not to depend on such features for "mission critical" things).

  -Amy

--
You received this message because you are subscribed to the Google Groups "Google App Engine" group.
To view this discussion on the web visit https://groups.google.com/d/msg/google-appengine/-/wPN-R3p3E-UJ.
To post to this group, send email to google-a...@googlegroups.com.
To unsubscribe from this group, send email to google-appengi...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/google-appengine?hl=en.