ViewTube

Infrastructure fails in ways too absurd to be fiction. Infra Pain turns real DevOps, SRE and sysadmin disasters into 60-second lessons: YAML betrayals, Friday deploys, 3am pages, Kubernetes doing exactly what you told it to (unfortunately). 

Every Short is a tiny postmortem, minus the blame, plus the jokes. New pain (almost) daily. Subtitles in English and Portuguese.

No production systems were harmed in the making of this channel. Several careers, maybe.

Shorts

The database was 9 hours behind. Everyone was "lying". 💀 #database

57 views

We deleted 800GB. The disk stayed full. 💀 #linux

120 views

Daylight saving charged every customer twice 💀 #devops

2 views

The load test hit production. It "worked". 💀 #sre

4 views

It's always DNS, part 2: the migration 💀 #devops

111 views

Nobody dared to reboot the 900-day server 💀 #sysadmin

56 views

One poison message took 4.2 million hostage 💀 #backend

55 views

We tested disaster recovery. The test caused the disaster 💀 #sre

60 views

50,000 customers got our "test test ignore" email 💀 #devops

35 views

2 million phones bricked by one certificate 💀 #mobile

51 views

The pager fired 340 times. The on-call phone was on Do Not Disturb. 💀 #oncall

66 views

We rolled back. The rollback made it worse. 💀 #devops

80 views

A Lambda triggered itself 2 million times. $11,000 in 40 minutes 💀 #aws

49 views

One service slowed down. Forty went down. Nothing crashed. 💀 #microservices

56 views

The migration was for staging. It ran on prod. 40 million rows. 💀 #database

56 views

10 pods. One got 100% of the traffic. The load balancer had a favorite 💀 #kubernetes

63 views

We deployed production to the wrong continent 💀 #aws

51 views

Every cache key expired at midnight. The database met 100k users at once. #sre

37 views

One semicolon broke 400 servers in 11 seconds 💀 #linux

289 views

A $12 renewal took the whole company offline 💀 #devops

49 views