36 ms·
Looks like a very cool solution if you can't use any of the common SaaS solutions out there like Pingdom or UptimeRobot. Making getting this thing reliable is
by tnolet 7y ago
Looks like a very cool solution if you can't use any of the common SaaS solutions out there like Pingdom or UptimeRobot.
Making getting this thing reliable is going to maybe a bit harder. Monitoring systems have this sort of "meta" issue of needing to be more reliable than the things they monitor. I'm building and running one as we speak and I had some interesting cases in that regard.
If you can us a SaaS tool, then here is my obligatory plug of my hosted / SaaS https://checklyhq.com https://checklyhq.com.
- sneak 7y ago> Monitoring systems have this sort of "meta" issue of needing to be more reliable than the things they monitor. Only slightly; they just need to have an extremely low likelihood of going down at the same time. It’s okay if your monitoring is down sporadically, just as long as it doesn’t overlap with the device under test also going down unexpectedly.
- tnolet 7y agocorrect, if you run a dedicated monitoring system for a handful of targets. If you run a SaaS or a very large monitoring system for a large corporation, this doesn't fly anymore. There is simply no moment available to be down. There is always overlap with some system.
- z3t4 7y agoJust set up a couple more monitoring servers. You want to test from many regions anyway to detect ISP shenanigans.
- runlevel1 7y agoTo an extent. A monitoring system that is unreliable risks losing trust. Monitoring systems that aren't trusted tend to be ripped out or ignored. If it is down too often, people will constantly question whether its silence is an indicator of health. If it triggers too many false alarms, people will constantly question whether an alarm indicates a real problem. If it gets bad enough, they might start ignoring them.
- brotandgames 7y agoCould you elaborate on the cases? This would be helpful for all of us.