Skip to content
Register

What's your acceptable downtime per year?

General Discussion by herebungkus 27 replies 3.8K views
herebungkus

Member OP

🌏 SEA represent

#1

I am curious, because I see many people here have different need. For my small web app, I think can tolerate maybe 1 day? But I am not sure if that is too much or too little. I use cheap VPS from RackNerd, so I expect sometimes slow-slow or restart.

Please vote:

  • Five nines (5.26 min/yr) — I run hospital system or something
  • Four nines (52.6 min/yr) — proper business, SLA with customer
  • Three nines (8.76 hrs/yr) — hobby but serious hobby
  • One nine (87.6 hrs/yr) — weekend maintenance fine
  • "If it down I just restart" — my current level wkwkwk

For me, I think three nines is dream, but one nine is reality. Thank you for your answer, everyone. Even if you disagree with me, thank you.

wrap it, ship it, pray it
PetraSuper

Member

network boot believer

#2

Hej, I run on server for 10 year, no SLA ever

"acceptable" depend on who pay. My malý server at home, if down, nobody care. No customer, no boss. I restart when I want.

But on config for work, we have four nines in contract. Hej, is different world. Money change everything.

I vote three nines for personal. No for business. Business need four nines or five nines, no discussion.

@herebungkus your one nine is 87 hour? That is almost 4 day. No, that is too much even for hobby. Sorry.

boot anything, anywhere, anytime
kat_fold

Moderator

Moderator

#3

Hey folks, friendly reminder: this is a poll thread, so let's keep it to voting and brief reasoning rather than debating each other's choices.

Herebungkus, thanks for framing this around SLO culture — it's a useful lens for the community.

I'll note that we have users ranging from students running tinaworm's legendary home lab to people managing production fleets. Both perspectives are valid here. No need to convince anyone their downtime tolerance is wrong for their context.

Carry on, and thanks for keeping it civil.

~be kind or be gone~
tinaworm

Member

#4
PetraSuper said:
I vote three nines for personal

Unacceptable. Even hobby infrastructure deserves discipline.

My home lab, 2021-present:

  • Dual WAN failover
  • UPS with automatic graceful shutdown
  • proxmox cluster, three nodes
  • Automated backups to secondary site
  • Measured downtime 2025: 4 minutes 12 seconds

That is not five nines. That is not four nines. That is effort without excuse.

Professionals and hobbyists are not different species. Only different budgets. Method remains.

-t

worst bandwidth, best stories
MI

Member

#5

One day is fine for a cheap vps lah

olespete

Member

Trust No One

#6

4 minutes measured, or 4 minutes noticed

airgapped, encrypted, faraday'd, still worried
lisbonred

Member

Tiny VPS Club

#7

Dual WAN is the real flex here

tinaworm

Member

#8
olespete said:
4 minutes measured, or 4 minutes noticed

Measured. Nagios + external uptime checker on Contabo VPS in Germany. If ICMP fails from both, clock starts. 4m12s was one event: power outage longer than UPS runtime, generator at secondary site did not start. I noticed 6 hours later when I checked email.

Noticed downtime is zero. I do not watch it breathe.

worst bandwidth, best stories
olespete

Member

Trust No One

1 #9

Contabo. There it is.

airgapped, encrypted, faraday'd, still worried
lisbonred

Member

Tiny VPS Club

#10
tinaworm said:
Generator at secondary site did not start

Secondary site has a generator. My secondary site is a USB stick in a drawer.

Different budgets indeed.

Post a reply

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

Post reply Preview Save draft