Is Contabo still taking new signups?

DNS resolution failure + auth stack down simultaneously smells like infra death spiral, not maintenance.

Check your registrar glue records — if they repointed nameservers without notice, your zone evaporated. Client area surviving on separate AS is classic zombification pattern.

`dig +trace yourdomain.tld` — post the NXDOMAIN chain.

Three days with zero substantive reply means their NOC is either outsourced to a chatbot or they've stopped paying staff. Either way, your client data's in limbo.

Start cold-migration planning now. 47 playbooks, you know the drill: rsync what you can reach, rebuild elsewhere, update glue before TTLs expire.

Contabo's been circling the drain since they stopped publishing status pages. This isn't a hiccup.

Three days radio silent during an outage? That's abandon-ship territory. Hope you've got dumps of anything that matters.

Stop treating this like a home router reboot. We're past traceroute theatre.

If the auth layer and resolution both vanished, either they torched their own netbox or the upstream pulled the route. Neither gets fixed by poking Cloudflare — https://www.cloudflare.com.

Contabo's ASN stopped announcing prefixes yesterday evening. That's not a DNS misconfiguration you can patch in a dashboard. That's a transit cut or a dead rack.

Anyone still waiting on "support" replies is burning daylight. Start rotating credentials you reused there and snapshot whatever you cached locally.

"machine died" + "attempting recovery" after 72h dark window = they had no DR plan

Bullet points:
- no hot spare: amateur hour
- "local backups you may have": blame shift
- single sysadmin: bus factor 1

`ansible` your exit strategy now

- name: evacuate
Hosts: production
Tasks:
- name: rsync to sane provider
Synchronize: src=/data/ dest=somewhere_else/

RTFM: "The SRE Book" — chapter 14, "accept no single points of failure"

Ser

Ngl that "machine died" line hit different. Single point of failure on a production box, zero replication, and they're making *you* bring the parachute?

I've seen this movie before. Had a RackNerd node pull the same stunt—turned out they were running consumer SSDs with no redundancy, just praying the SMART stats stayed green.

48h recovery window with no status page updates, then "attempting data recovery" as a closing argument. That's not a hosting operation, that's a coin flip with your datasets.

If they actually had RAID or any sane backup cadence, they'd be spinning up a replacement already, not panning for gold on a dead platter.