I check HTTP status code and response time, not just ICMP. If it is 200 but slow, I mark degraded. If it is 502/503, down. ICMP is useless for actual service health. My KnownHost checks hit a small PHP page I asked them to put on each location. They refused. So I check their main website. Not the same but closer than ping.
Your status page is always green
I am in Vilnius. We have a local VPS provider with no status page at all. Just a Telegram channel where the owner posts "router firmware update tonight" and "back online, sorry for delay." It is more useful than any green dashboard I have used.
In Lima we do not have many options. I use AWS Lightsail because at least when they have issues, their status page eventually turns yellow. It might take 45 minutes but it happens. The "eventually honest" bar is low and most do not clear it.
AWS is not special. I watched their status page stay green during a us-east-1 DNS issue that broke half the internet. The page updated 90 minutes after Hacker News frontpaged it. "Eventually honest" is just "caught by too many people to ignore."
speedasad said:
"Eventually honest" is just "caught by too many people to ignore."
This is the business model. Status pages are not for customers. They are for PR. Green until social media forces otherwise. Then yellow with "increased error rates" instead of "we broke it." I have automated my own with Nagios + Prometheus, four external VPS locations, published to GitHub Pages. Cost: 12€/month. Trust: still imperfect, but at least I know where the blind spots are.
nate_pad said:
I was not sure if I was being dumb
You are not dumb. The system is designed to make you doubt yourself. Green page, check your router, check your DNS, check your ISP. By the time you prove it is them, the incident is over and the page still says green. The dumb thing is trusting it. The smart thing is what Azhar763 did: build your own, make it public, let them explain the gap.