Performance issues on latest PublicLab.org, resolved, notes

[Jeffrey Warren]

Hi, all -- 

There were some performance issues yesterday which we resolved last night after some great sysadmin work by Sebastian. Its all good and the site is running smoothly now. 

This is not a big deal and is to be expected with big new code publication from time to time, although we should try hard to look through code with a sharp eye (myself included! I missed this when I reviewed PR code)

I ended up commenting out the UsersController#map method, and this line (and a similar one just below it) seem to have caused the problems we had:

https://github.com/publiclab/plots2/commit/abf4ba8017f94c5efea52ad7c13d35aa68434e0a#diff-4e05ad0d64e6100656b63ad1e78f32c5R306

That's a call of DrupalUsers.all -- a table that has hundreds of thousands of records; actually mostly spam users. It was called occasionally through the day and (with so many results) filled up memory on the server, causing outages. That action was triggered by visiting publiclab.org/users/map, which now returns not much, since I disabled it. (i disabled the functional tests for that action too).

Just be aware how many records you're going to get on a given call -- the tags table is also quite full, as is node_revisions (which includes large text blobs). 

Lalith, you're doing amazing work, so don't worry about this -- i'm really just making a general note to everyone to keep an eye on resource usage on the production database. 

Thanks, all! Very excited to have all this new code running!

Jeff