max_wait_mins seems to break stuff

[stevenpollack]

location=Berlin&max_wait_mins=58 is just before things break (55.5MB return).

Heroku logs:

2016-04-16T22:04:10.727076+00:00 heroku[router]: at=error code=H12 desc="Request timeout" method=GET path="/double-dips?location=Berlin&days_from_now=1&max_wait_mins=60" host=dairy-queen.herokuapp.com request_id=9536b7cb-5eb0-4ba6-9651-a6f841b75ee0 fwd="91.64.93.249" dyno=web.1 connect=0ms service=30000ms status=503 bytes=0
2016-04-16T22:04:11.153126+00:00 app[web.1]: [2016-04-16 22:04:11 +0000] [3] [CRITICAL] WORKER TIMEOUT (pid:7)
2016-04-16T22:04:11.155690+00:00 app[web.1]: [2016-04-16 22:04:11 +0000] [7] [INFO] Worker exiting (pid: 7)
2016-04-16T22:04:11.407705+00:00 app[web.1]: [2016-04-16 22:04:11 +0000] [17] [INFO] Booting worker with pid: 17
2016-04-16T22:04:14.931354+00:00 heroku[router]: at=error code=H12 desc="Request timeout" method=GET path="/double-dips?location=Berlin&days_from_now=1&max_wait_mins=60" host=dairy-queen.herokuapp.com request_id=34a48476-51c1-4b33-b7cb-c439643c405f fwd="91.64.93.249" dyno=web.1 connect=0ms service=30001ms status=503 bytes=0
2016-04-16T22:04:15.481444+00:00 app[web.1]: [2016-04-16 22:04:15 +0000] [3] [CRITICAL] WORKER TIMEOUT (pid:15)
2016-04-16T22:04:15.484211+00:00 app[web.1]: [2016-04-16 22:04:15 +0000] [15] [INFO] Worker exiting (pid: 15)
2016-04-16T22:04:15.772639+00:00 app[web.1]: [2016-04-16 22:04:15 +0000] [19] [INFO] Booting worker with pid: 19
2016-04-16T22:04:16.778719+00:00 heroku[router]: at=error code=H12 desc="Request timeout" method=GET path="/double-dips?location=Berlin&days_from_now=1&max_wait_mins=60" host=dairy-queen.herokuapp.com request_id=4710c800-5a02-4e8e-8b78-17fa3ddaa413 fwd="91.64.93.249" dyno=web.1 connect=0ms service=30001ms status=503 bytes=0
2016-04-16T22:04:02+00:00 app[heroku-redis]: source=REDIS sample#active-connections=1 sample#load-avg-1m=0.04 sample#load-avg-5m=0.095 sample#load-avg-15m=0.105 sample#read-iops=0 sample#write-iops=0 sample#memory-total=15666160.0kB sample#memory-free=13341796.0kB sample#memory-cached=555688kB sample#memory-redis=294048bytes sample#hit-rate=1 sample#evicted-keys=0
2016-04-16T22:04:30.439691+00:00 heroku[router]: at=error code=H12 desc="Request timeout" method=GET path="/double-dips?location=Berlin&days_from_now=1&max_wait_mins=60" host=dairy-queen.herokuapp.com request_id=2ef6baf4-813a-412c-b336-aea7ea8bbe31 fwd="91.64.93.249" dyno=web.1 connect=0ms service=30000ms status=503 bytes=0

[stevenpollack]

This doesn't seem to break for Hinsdale, MT, which means this may be a size/scale issue.

[stevenpollack]

same issue with ?location=Paris&max_wait_mins=11 (53.4MB return) and error at max_wait_mins=12...

[stevenpollack]

re-ran the berlin query locally... True return is 71.0MB, which means this is heroku's dynamo that's dying (probably insufficient memory). The memory usage for the app is out of control. Will have to look into streaming returns...

http://flask.pocoo.org/docs/0.10/patterns/streaming/
https://devcenter.heroku.com/articles/python-rq

[gabericharde]

Oh boy, well sorry for those few big Berlin hits I did. I am using Lexington, KY & Lexington, VA but can start using Hinsdale, MT exclusively to limit the size of the return.

[stevenpollack]

It's fine. This'll get fixed shortly with an "include" parameter to let the
user filter the theatres.

I would implement streaming but I was convinced that this wouldn't actually
address the core use-case. Not to mention it's unclear how a front end
would make use of a streaming return.
On Mo., 18. Apr. 2016 at 10:23, Gabriel RiCharde [email protected]
wrote:

Oh boy, well sorry for those few big Berlin hits I did. I am using
Lexington, KY & Lexington, VA but can start using Hinsdale, MT exclusively
to limit the size of the return.

—
You are receiving this because you authored the thread.
Reply to this email directly or view it on GitHub
#10 (comment)