OVER the coming days we will change some configs. It ought not take more than 1 hour in total, some time between 1PM and 6PM GMT on Thursday. IRC will be up as usual (even though I myself will be disconnected in the interim) and the Gemini capsule will be offine. The Web site will work as usual, whereas IPFS is distributed so it will probably cope in the absence of my node. This isn't an accident or incident; we're getting our home connection upgraded to optical fiber (a.k.a. "fiber optics"), which ought to help us in a lot of ways, especially the self-hosting and decentralisation drive. We have everything ready for the installation, including the time slot and shiny new equipment. A few months ago such downtime wouldn't have been a big deal, but nowadays we server over 100,000 Gemini pages per month, so even one hour of downtime certainly counts. I have one screen devoted only to Gemini (ensuring uptime and checking load). We've suffered a number of DDOS attacks lately. Many of the things we cover/report on are very unique, so if/when we're offline information is being suppressed. A multitude of protocols and hosts (even for IRC) should bring up closer to 100% uptime. ⬆
Improving and ensuring uptime by 24/7 monitoring (load, disk space, DDOS patterns etc.)