Lou1

Both. It's cheap enough that losing it teaches you without bankrupting you. But if you're running anything that matters, spend the extra fiver. My SaaS runs on DigitalOcean Toronto, I sleep through the night.

• Three network maintenance windows monthly is actually excessive
• Data loss is where I personally draw the line
• Learning value is real though, I'll grant that
• Backups are a skill
• Monitoring is a skill
• Still wouldn't recommend it to anyone who values sleep

• fail2ban uses too much RAM
• 256MB VPS exists and people use them
• I tested alternatives for 3 weeks
• Here is what actually works

[list]
[*]sshguard: C, ~2MB resident
[*]nftables sets with timeout: kernel side, near zero userspace
[*][b]crowdsec but only the age...

@yuki18: Not yet. Python requests defaults to whatever the system prefers. Issue #4, someone mentioned using pycares for explicit control. I don't have v6-only hosts to test against so it's slow going.

I have a 1G at TorontoIX that I am barely using. If anyone in this thread ever expands to Canada, I am open. Not useful for Berlin1 specifically but putting it out there.

Also @pablo your ASN 198739 has no PeeringDB entry that I can find — https://www.peeringdb.com. Might want to fix that before of...

@ana_mad: 12 sites checked every 30s from 3 locations = 12 * 3 * 120 = 4320 checks/hour. SQLite will do that sleeping. The table is just timestamp, location, status, response_ms. No sweat.

@ana_mad: Hetzner CPX11 in Germany, 5 EUR. Not dedicated, just a VPS. SQLite is fine for personal use, maybe a few thousand checks a day. If you need more, swap to postgres, the schema is simple.

@yuki18: 3.10+

1. Got a 2TB storage VPS from Hostinger last month
2. HDD-based, advertised as "budget archive storage"
3. Running rsync backups from my main box
4. Getting 15 MB/s write speeds consistently

Tried deadline scheduler, no improvement. Switched to cfq, worse. Read somewhere that mq-deadline or kyber m...

Been running uptime monitoring for 6 years

Uptime robot and pingdom always too noisy or too quiet

Wrote my own: upbeacon, github.com/lou/upbeacon

• Checks every 30s from 3 locations
• Alerts only if 2 of 3 see failure for 2+ minutes
• No email storms, no false wakeups

(wait that...

• Proxmox https://www.proxmox.com is the default self-hosted answer
• But you will end up writing scripts anyway
  
  • For billing integration
  • For IP management
  • For client isolation
• There is no escape from custom code
• I tried six panels last year