Contabo's 72-hour 'network upgrade' and silence

Contabo announced a 'routine network upgrade' for January 15. I expected 2 hours. We got 72.

Timeline:
- Hour 0: Maintenance window starts. No status page update.
- Hour 6: My monitoring goes red. Error budget for quarter: exhausted.
- Hour 18: First acknowledgment. 'Unexpected complications.'
- Hour 48: 'Engineers are working diligently.' No ETA. No transparency.
- Hour 72: Service restored. No postmortem offered.

I run blameless postmortems at my day job. I believe in transparency. Contabo provided none. This is not about the outage itself—hardware fails, upgrades go wrong. This is about the communication vacuum. When you have customers running production workloads, silence is not a strategy. I have since migrated my critical path to Vultr (https://www.vultr.com). I kept the Contabo VPS for non-essential workloads only. Transparency builds trust. Silence destroys it.

# my "fun little project" during the outage
# wireguard tunnel through a €3 NAT VPS at Hetzner
# routed all Contabo traffic through it
# latency was terrible but it actually worked

[Interface]
PrivateKey = REDACTED
Address = 10.200.200.2/24
ListenPort = 51820

[Peer]
PublicKey = REDACTED
AllowedIPs = 0.0.0.0/0
Endpoint = hetzner.com:51820

Backup plans, folks. Always have backup plans.

Srelet (fellow sufferer (though (to be fair (I suppose)) my 'upgrade' was at RackNerd (which is to say (not Contabo (but (and this is the point ))) the pattern (the pattern (!)) repeats across the industry))) I got six hours of 'we are improving your experience' (improving (!)) followed by three days of radio silence (radio (as in (nothing (absolutely nothing)))) and then (the cherry (the maraschino cherry)) a credit of (wait for it) one month on a (wait more) yearly plan