Skip to content

What do you use for monitoring without another VPS?

VPS Hosting by annexbi 11 replies 703 views
#1

So here's the thing (and I know (I KNOW) someone will tell me this is stupid) I have one lonely Contabo KVM in RackNerd's cheapest tier (don't ask (or do (I love pain apparently))) and I want to know when it dies without paying for a second VPS to watch the first one. UptimeRobot free tier exists but then you're just... trusting them? Which is the same problem with extra steps. Self-hosted healthchecks.io (which I tried (briefly (very briefly))) needs somewhere to live, which brings us back to square one. Local cron pinging a Matrix bot I run on my home Pi. It just works. (It doesn't. It will fail the exact moment my power goes out, which is when I actually need to know.) Anyone solved this paradox without just giving Oracle more of your soul?

push. done. coffee.
#2

Your Pi solution reminds me of the old smokeping on Soekris net4501 days. Ugly, reliable, nobody trusted it but it outlived three "cloud" migrations.

#3

Wesh anne ! Le server حلو but le monitoring ... incha'allah the uptime, non ?

Moi j'ai fait le cron sur le router (OpenWrt) qui envoie vers un groupe Matrix, mais c'est la même merde : quand le router tombe, personne sait rien . Le vrai problème c'est le paradoxe du watcher qui watch rien .

RAID 1: because paranoia pays
#4
annexbi said:
It will fail the exact moment my power goes out, which is when I actually need to know.) Anyone solved this paradox without just giving...

Obviously you have solved this perfectly, or whether you have not. How do you say... the Pi will save you. Clearly. /s

I run a second Pi at my mother's house, which is obviously a robust enterprise architecture. She thinks it's a "weather thing." It has not been plugged in since March. /s

#5

Bro . I use very cheap oracle free tier gan . Just for ping only .

But ya ya same problem bro . Oracle also can die . I just accept lor . Check when check can already . No perfect thing gan ...

#6

There was same paradox with my watcher setup on OpenWrt

#7
moebig said:
Le vrai problème c'est le paradoxe du watcher qui watch rien

This is the only true thing said so far. The rest is cope.

I will tell you what I actually do because I am not twelve. I pay for UptimeRobot Pro. 15 USD per year. It is not the same as self-hosted because their entire business is being up, and they have pages of SLA. My Contabo in Nuremberg could be on fire and I would still get the SMS.

You want free? Accept that free has this paradox. You want reliable? Pay for reliable. There is no third door.

#8
Sven3 said:
I pay for UptimeRobot Pro. 15 USD per year.

I did not know the Pro was that cheap. I was looking at their enterprise page like an idiot.

Still. Still. "Their entire business is being up" -- so was Pingdom's, and I remember that acquisition. Not saying UptimeRobot sells tomorrow but the principle bothers me.

push. done. coffee.
9 #9
Sven3 said:
Pay for reliable

Or run six Oracle free tier instances across three regions and have them ping each other in a ring. When three agree the fourth is dead, you get the mail. I have done this for fourteen months. Zero false positives, one true positive when Seoul actually went down.

Is it stupid? Extremely. Does it cost zero? Yes.

licensing is a suggestion
#10
olya said:
Six Oracle free tier instances

Bro your time also cost money lor . Six instance to maintain just for ping . My cheap oracle one already enough headache gan .

Post a reply

You need an account to reply. Log in or register to join the conversation.

Post reply Preview Save draft