Repository navigation
Synchronous Mode #101
Description
I think both #34 and #51 are symptoms of the same underlying issue - that our threading does not play nicely with Celery and Google App Engine.
As @calvinfo mentioned in #51 (comment); the right solution would be to allow users to control/skip the async nature of this library and use it as a dumb HTTP API wrapper when needed.
Reacted by Jeff Kerr and Josh Smeaton
Activity
f2prateek commentedon Nov 22, 2017
Two potential paths we can take here:
russlevy commentedon Mar 19, 2018
Synchornous mode would be great for me as well.
samcheng commentedon Mar 19, 2018
savraj commentedon Mar 19, 2018
the solution for google app engine (for me at least) is to use the built in task queue and just use the http api with urlfetch
poswald commentedon Mar 20, 2018
That pull looks good but Is there going to be a way to configure/explicitly specify a
SyncConsumereven if you're not in AWS Lambda or Google App Engine?samcheng commentedon Mar 20, 2018
#105 works for me. It bypasses the Consumer class entirely. I've been using it for backfills, which appear to exacerbate the Celery issue. We're on Heroku instead of GAE, for what it's worth.
gojefferson commentedon Feb 20, 2019
Synchronous mode would be a great feature. I'm having problems using analytics with django-rq, and I think the problem would be solved if I could make synchronous calls when I want to.
Add sync_mode option (#147)