ronwit said:
Don't overengineer
This. Following.
Budget: $20 total. One big VPS or four tiny ones?
I have 8 VPS across 4 providers. The monitoring overhead alone is non-trivial. Four OVHcloud boxes means four sets of kernel updates, four SSL renewals to track, four fail2ban logs. Automation helps but you still get paged. One box = one pager.
Ansible exists.
AsadiCoder said:
Ansible exists
Ansible doesn't make network partitions disappear. Or DNS propagation delays. Or the fact that your four "redundant" boxes are probably in same Gravelines DC on same power feed.
ben20
Member
Fly7 said:
Same Gravelines DC on same power feed
OVHcloud VPS are in different availability zones within region. I have documented this. But AZ is not datacenter, it's logical separation. The SBG2 fire proved that.
SBG2 was Strasbourg dedicated servers, not VPS. Different failure mode.
The fire destroyed the building. VPS in other SBG zones were affected by smoke and power, not flames. I was there, I had boxes down for 48 hours.
I have VPS in OVHcloud Gravelines. Wait, no I don't, I have them in Gravelines and Warsaw. The point is: check where your stuff actually is.
norddude said:
Check where your stuff actually is
This. I thought my "France" VPS was in Paris, turns out Gravelines. Not same thing.
For $20 you could also do 2x 2GB OVHcloud and have simpler bot placement + some redundancy. Middle path.