Hey, not sure what your Solr index size is or what your user count is at now, but that Heroku Solr cost seems fairly large for this side project.
You might want to consider looking at Solr on Google Cloud. With a $300 credit from them, I've only paid cents for the past month or two, but I haven't launched to a lot of traffic yet.
Mostly just letting you know that Solr on GC is possible and maybe it could save you a little cash.
I'm not using the Heroku Solr, aside from the pricing they don't seem to let you use a custom data model, whereas I just wanted to zip the index and upload it.
I have a VM running at Linode, which is unfortunately a different datacenter, but it only added about 25ms to the result times. Google Cloud is an interesting option though!
Ah, gotcha, I was going off your Donation page on findlectures.com that mentioned Heroku Solr at $50/month.
So far so good with Google Cloud and Solr. There's a Bitnami Solr image that I used to get it all set up and with that credit I haven't paid anything yet (though I don't have your traffic of course).
Cool project by the way. :)
EDIT: Oh, looks like another article on dev.to talks about GC vs AWS so that might be handy for you: dev.to/bugfenderapp/google-cloud-p... if you haven't looked into it (not that you're using AWS for this, but insight anyway).
Memory use has been the big issue, rather than traffic. It seems like solr uses RAM proportional to disk space to run well (1-2x maybe?), so I've hit limits as I grow the index, but I'm still figuring that out.
Once I made it so you can search closed captions, and see what time in the video matched, the index started getting bigger a lot faster (that is timings + caption, and a non-timed version of the transcript).
For further actions, you may consider blocking this person and/or reporting abuse
We're a place where coders share, stay up-to-date and grow their careers.
Hey, not sure what your Solr index size is or what your user count is at now, but that Heroku Solr cost seems fairly large for this side project.
You might want to consider looking at Solr on Google Cloud. With a $300 credit from them, I've only paid cents for the past month or two, but I haven't launched to a lot of traffic yet.
Mostly just letting you know that Solr on GC is possible and maybe it could save you a little cash.
I'm not using the Heroku Solr, aside from the pricing they don't seem to let you use a custom data model, whereas I just wanted to zip the index and upload it.
I have a VM running at Linode, which is unfortunately a different datacenter, but it only added about 25ms to the result times. Google Cloud is an interesting option though!
Ah, gotcha, I was going off your Donation page on findlectures.com that mentioned Heroku Solr at $50/month.
So far so good with Google Cloud and Solr. There's a Bitnami Solr image that I used to get it all set up and with that credit I haven't paid anything yet (though I don't have your traffic of course).
Cool project by the way. :)
EDIT: Oh, looks like another article on dev.to talks about GC vs AWS so that might be handy for you: dev.to/bugfenderapp/google-cloud-p... if you haven't looked into it (not that you're using AWS for this, but insight anyway).
Good point, I should update the donation page!
Memory use has been the big issue, rather than traffic. It seems like solr uses RAM proportional to disk space to run well (1-2x maybe?), so I've hit limits as I grow the index, but I'm still figuring that out.
Once I made it so you can search closed captions, and see what time in the video matched, the index started getting bigger a lot faster (that is timings + caption, and a non-timed version of the transcript).