jay...@basshero.org
unread,Oct 11, 2012, 4:38:32 AM10/11/12Sign in to reply to author
Sign in to forward
You do not have permission to delete messages in this group
Either email addresses are anonymous for this group or you need the view member email addresses permission to view the original message
to diaspo...@googlegroups.com
Hi guys and gals,
Currently our team is running a Diaspora* pod for our internal social media needs (pretty awesome :)). At first I just demoed it with a free Heroku pod (1 web dyno with embedded resque worker) and now my boss is super excited :)
It's very slow unfortunately, the part where it loads the stream, with little or no traffic taking up to 3 seconds to retrieve the stream. Our database was at first externally but then as it was really slow we changed it to Xeround which an app for Heroku and is located in the Amazon US East location. That solved general page loading which is now quite fast - but still the stream is really slow.
We tried putting a separate worker (1 web, 1 worker), changing the following config "unicorn: embed_resque_worker: false" and adding a worker line in Procfile.
It didn't really make any difference, so we tried also upping web dynos up to 3 - which made absolute no difference in stream loading. From the logs I could see that requests were going to three dyno's.
What are we doing wrong here? Anyone who runs a non-free nice performing pod on Heroku willing to share tips on how to configure it? Could the issue really be just network lag?
The only thing that changed with upping the dynos to three was getting some MySQL too many connections errors - could that be related to the slowness, the Xeround MySQL plan not being able to offer data in enough connections?
My boss wants to use Heroku as he doesn't want me spending time upkeeping a VPS server ;) So I need to get it to run faster otherwise the team members might want to look at other options :P Anyone know any options besides Heroku in European data centers?
Any tips welcome! Thanks!
Br,
Jason