I'll start! I had a OVHcloud VPS in their London location that started dropping packets every evening like clockwork. 20% loss from 20:00-23:00 UTC. I opened a ticket with mtr logs, iperf3, the works.
47 messages later — forty-seven! — their final reply: "The issue is your system timezone is set to EST. Their network maintenance window is UTC. Please correct your timezone and close the ticket."
My timezone. Caused their packet loss. I asked how my /etc/localtime affected their upstream router. They escalated to "senior network engineer" who said: "The customer does not understand how time synchronization affects routing protocols."
I run NTP. My clock was fine. I moved to Contabo the next week. Insane value on their Ampere Altra line, runs circles around OVHcloud's old Xeons. 80 cores for the price of 8!
Specs of what I moved to:
- Ampere Altra M96-28
- 256GB DDR4 ECC
- 2x3.84TB NVMe (RAID-1)
- 10Gbps unmetered
What's your worst?