Could anyone offer any tips on ways to optimize the performance of filtered continuous pull replications? I'm experiencing some slowness pulling data from my central couchdb to mobile clients. I'm guessing part of the problem is that my filter contains string manipulation logic (substring comparisons) which I need to fix. But I'm hoping there are some other things we can try to improve the performance? Are there any couchdb configuration settings people would recommend experimenting with? Would setting nodelay=true make a difference? I really appreciate any advice people can offer!
Thanks! Bill
