Reply to thread

Message

<blockquote data-quote="Phil Graham" data-source="post: 47473" data-attributes="member: 430">Re: More on donations, costs, and the futureI would be very disappointed if the current server infrastructure and our software architecture had a problem at 2x the current user base, but the bottlenecks that creep up as pages served grows are sometimes very subtle and we need to be prepared.For this transition we spent our money moving from Behringer-grade hosting to Clair Global-grade hosting. Already that has paid dividends. As a concrete example, there is a bug in how our distribution of Linux cleans up PHP old processes that would have crashed the server on Monday or Tuesday had it not been for the ability of our hosting to dynamically give us more CPU time when needed. We were spiking to almost 200% of allotted cpu usage every half hour when the errant process ran. This extra cpu headroom let us track down and fix the bug. Without features like this we would already have been down!To give everyone a next tier of donations to shoot for, we really need to pay for Google, Cloudflare, and Pingdom. All three of these services are critical to how the site architecture is set up. Google makes our mail reliable, Cloudflare keeps the spam away while reducing server bandwidth, and Pingdom keeps our uptime high. An extra $90/month in donations would cover these three pieces, and solidify the ongoing core architecture.Eat the elephant one bite at a time,-Phil</blockquote>

[QUOTE="Phil Graham, post: 47473, member: 430"] Re: More on donations, costs, and the future I would be very disappointed if the current server infrastructure and our software architecture had a problem at 2x the current user base, but the bottlenecks that creep up as pages served grows are sometimes very subtle and we need to be prepared. For this transition we spent our money moving from Behringer-grade hosting to Clair Global-grade hosting. Already that has paid dividends. As a concrete example, there is a bug in how our distribution of Linux cleans up PHP old processes that would have crashed the server on Monday or Tuesday had it not been for the ability of our hosting to dynamically give us more CPU time when needed. We were spiking to almost 200% of allotted cpu usage every half hour when the errant process ran. This extra cpu headroom let us track down and fix the bug. [I]Without features like this we would already have been down![/I] To give everyone a next tier of donations to shoot for, we [I]really[/I] need to pay for Google, Cloudflare, and Pingdom. All three of these services are [I]critical[/I] to how the site architecture is set up. Google makes our mail reliable, Cloudflare keeps the spam away while reducing server bandwidth, and Pingdom keeps our uptime high. An extra $90/month in donations would cover these three pieces, and solidify the ongoing core architecture. Eat the elephant one bite at a time, -Phil [/QUOTE]

Verification