It does HTTP with keyword check, which is useful. I verify my login page has "Sign In" not just 200 OK.
But it only runs from LA. Good for checking your global CDN, bad for finding regional blackholes.
It does HTTP with keyword check, which is useful. I verify my login page has "Sign In" not just 200 OK.
But it only runs from LA. Good for checking your global CDN, bad for finding regional blackholes.
I use HetrixTools for the external checks. $15/year for 15 monitors from multiple locations. Not self-hosted, not fancy, but it caught a DNS issue my internal stuff missed.
Worth considering if you want cheap and do not need to own the whole pipe.
HetrixTools is US-based. I am not comfortable with my domain data flowing there. Datenschutz first.
I use statuscake from within EU. More expensive, but the data stays where the law protects it.
I worked at an ISP and we used SmokePing for latency trends. RRD graphs, ugly as sin, but you see degradation before it becomes outage.
Not for everyone. Needs a server, needs patience. But the historical data saved us from blaming the wrong upstream three times.
What about certificate monitoring? That is not exactly uptime but it is domain-related failure. I have had customers call because the site "does not work" when it is just expired cert.
I use a free certbot hook that posts to Slack. Simple, effective, zero extra cost.
Certificate I Am Checking Separate Also Because The Auto Renew It Is Failing Sometimes
The Letsencrypt It Is Not Always Working If The Firewall Is Too Strict
I Am Thinking The Monitoring Must Check The Certificate Even If The HTTP It Is Up
My terrible bash script checks cert expiry with openssl s_client!!! It is three lines!!! It has never failed because it is too stupid to fail!!!
Actually it failed once when openssl changed output format but that is ONCE in four years!!!
Certificate checks, multiple locations, different networks, Telegram alerts. I have my homework.
One last thing: how often? Every 60 seconds seems like overkill for a small site. Every 5 minutes? 15?
I do 60 seconds on HTTP from external, 5 minutes on DNS and cert. The 6-hour outage you had? With 60-second checks the worst case is you know in 60 seconds.
For $20 and one domain, I would do 1 minute external, 5 minute internal. You are not polling a thousand hosts.