2 ms·

Can you reduce your server costs? At around 0.16K per request * 5,000,000 requests = 800,000K (781.25M) * 30 days = about 23.5G per month. That is well under t

Can you reduce your server costs? At around 0.16K per request * 5,000,000 requests = 800,000K (781.25M) * 30 days = about 23.5G per month. That is well under the 1TB you'd get per month for a $5 VPS at some place like https://www.digitalocean.com https://www.digitalocean.com.

geuis 13y ago
Thanks Matt. I'm pretty set on Linode though. Great service and reliability. Also, I don't have the free time right now to completely rebuild a server instance and do migrations.
- matt_heimer 13y ago
  Probably doesn't help but knowing how much you pay is a problem for me, given that it is such a simple service and how much you are paying I feel like it'd be better if I spent the 2 hours to re-implement it and setup my own VPS. 2 hours + 50 dollars == 10 months of service or 50 dollars == 1 month of service. I have a feeling that a lot of the HN crowd would use your service if it was around but we all have hosting solutions we can use to re-implement it so honestly it becomes a choice of giving you money or spending two hours (and it looks like another commenter already implemented a clone). It seems like more like a value-add to another service like Pingdom than something that will ever be self sustaining (this coming from someone that runs a few out of pocket things for people that will never be self sustaining). To be blunt, I doubt you'll break even and if you are having trouble affording keeping the server up you should probably kill it. If want to give something to the community that doesn't cost money you could always release the server impl on github. But if you still want to give it a go and seeing as you weren't getting much in the way of answers to your question (my answer included) I searched and it seems the same question got asked on HN recently and there were some more useful links: https://news.ycombinator.com/item?id=6113484 https://news.ycombinator.com/item?id=6113484 From everything I've heard I'd try Stripe first (no first hand exp myself)
  - geuis 13y ago
    I've run jsonip.com for years. It supports millions of requests a day. More than anyone else, its organic growth and scale has surprised me. The server I have costs $50. That provides me with enough capability to support jsonip and other sites. There's not a chance in hell I'm going to kill it. That being said, things have been tight for me recently and I'm trying to get this straightened out. But I figured it would be worth the 10 minute effort to see if any of the hundreds-thousands of people that use the service on some very, very high traffic sites were willing to chip in a bit since its been free, and will remain so indefinitely.
    - matt_heimer 13y ago
      But you aren't checking to if your users are willing to support it and honestly you have almost no way to reach your users. Your users have no reason to visit your site, most probably don't even read the about page linked to in your output. If you do implement donations you need to some way more than just adding a donation link in the output or to the about page. I'd suggest trying to drive some traffic to your human readable site once you have donations setup. The only people that will probably care enough to visit your website and see a donation button are the people that are about to use your service for the first time. Unless they are just stumbling across your URL in some stackoverflow post and then you'll probably never get a chance to connect with them. Create several tutorial pages on how to call your service. Create a library for people that don't know JavaScript enough to use XHR. Make a library with XDomainRequest support so IE 8 & 9 will work with CORS. Make a jQuery plugin, etc, etc. Don't do it as blog postings, make a real tutorial section. Every single code example you create should embed the donation URL as comment.
      - geuis 13y ago
        These are actually good, practical examples. Thanks.

geuis 13y ago

Thanks Matt. I'm pretty set on Linode though. Great service and reliability. Also, I don't have the free time right now to completely rebuild a server instance and do migrations.

matt_heimer 13y ago
Probably doesn't help but knowing how much you pay is a problem for me, given that it is such a simple service and how much you are paying I feel like it'd be better if I spent the 2 hours to re-implement it and setup my own VPS. 2 hours + 50 dollars == 10 months of service or 50 dollars == 1 month of service. I have a feeling that a lot of the HN crowd would use your service if it was around but we all have hosting solutions we can use to re-implement it so honestly it becomes a choice of giving you money or spending two hours (and it looks like another commenter already implemented a clone). It seems like more like a value-add to another service like Pingdom than something that will ever be self sustaining (this coming from someone that runs a few out of pocket things for people that will never be self sustaining). To be blunt, I doubt you'll break even and if you are having trouble affording keeping the server up you should probably kill it. If want to give something to the community that doesn't cost money you could always release the server impl on github. But if you still want to give it a go and seeing as you weren't getting much in the way of answers to your question (my answer included) I searched and it seems the same question got asked on HN recently and there were some more useful links: https://news.ycombinator.com/item?id=6113484 https://news.ycombinator.com/item?id=6113484 From everything I've heard I'd try Stripe first (no first hand exp myself)
- geuis 13y ago
  I've run jsonip.com for years. It supports millions of requests a day. More than anyone else, its organic growth and scale has surprised me. The server I have costs $50. That provides me with enough capability to support jsonip and other sites. There's not a chance in hell I'm going to kill it. That being said, things have been tight for me recently and I'm trying to get this straightened out. But I figured it would be worth the 10 minute effort to see if any of the hundreds-thousands of people that use the service on some very, very high traffic sites were willing to chip in a bit since its been free, and will remain so indefinitely.
  - matt_heimer 13y ago
    But you aren't checking to if your users are willing to support it and honestly you have almost no way to reach your users. Your users have no reason to visit your site, most probably don't even read the about page linked to in your output. If you do implement donations you need to some way more than just adding a donation link in the output or to the about page. I'd suggest trying to drive some traffic to your human readable site once you have donations setup. The only people that will probably care enough to visit your website and see a donation button are the people that are about to use your service for the first time. Unless they are just stumbling across your URL in some stackoverflow post and then you'll probably never get a chance to connect with them. Create several tutorial pages on how to call your service. Create a library for people that don't know JavaScript enough to use XHR. Make a library with XDomainRequest support so IE 8 & 9 will work with CORS. Make a jQuery plugin, etc, etc. Don't do it as blog postings, make a real tutorial section. Every single code example you create should embed the donation URL as comment.
    - geuis 13y ago
      These are actually good, practical examples. Thanks.