4 ms·

Yes, moving to VM is definitely doable, but now, being a 1 person dev team, it is challenging in maintenance. My fear is, being on a vacation, and suddenly th

• rozenmd 4y ago
  Self healing VMs - look into fly.io, they just restart when out of memory etc took about a day to rewrite and a week of 30 mins per day to figure out how to optimise my code for self healing. knowing my bill is capped by number of VMs * memory, saves a lot of stress
  • rozenmd 4y ago
    In case you're curious, I wrote about my experience here: https://onlineornot.com/on-moving-million-uptime-checks-onto-fly-io https://onlineornot.com/on-moving-million-uptime-checks-onto...
• amerkhalid 4y ago
  Back in days when our services ran on metal or VM, if we got paged during off-hours and had something else going on, a simple reboot almost always fixed the issue. We, developers, never liked the reboots though, always wanted to find out root cause, so that we won't be paged again. So, I guess, we moved to the cloud. Now we don't get paged in the middle of night. But yeah if it was my own company and I was on vacation, I rather take 30 seconds to reboot the server instead of worry about paying thousands of dollars. As for lack of network coverage, you could do scheduled daily/hourly reboots while on vacation, if it makes sense for your service. Or if your an outage will cause massive disruption to your users, then perhaps hire a part-time sysadmin while on a vacation.