The Friday 5 PM Production Deploy
Description
A popular meme featuring a grey and white tabby cat walking confidently and calmly towards the camera. In the background, a massive plume of grey smoke billows up from a hill or mountain, indicating a large explosion or volcanic eruption. The setting is a field of dry, brownish grass under a partly cloudy blue sky. Below the image, a caption reads: 'Leaving work after deploying to production on Friday at 5pm'. An Instagram watermark for '@alpha.coder' is visible in the bottom right. This meme format, often called 'Cool Cat Walking Away from Explosion,' humorously juxtaposes a nonchalant character with a chaotic event they seemingly caused. The technical joke centers on the widely recognized anti-pattern of deploying software to a live production environment late on a Friday. This practice is notoriously risky because if any critical bugs or system failures occur, the development team is likely unavailable over the weekend to address them, leading to prolonged outages and frantic on-call pages. The meme perfectly captures the reckless or blissfully ignorant attitude of a developer causing potential weekend chaos and then promptly leaving, resonating with any engineer who has experienced the fallout from such a decision
Comments
8Comment deleted
A junior dev pushes to prod on Friday because they don't know any better. A senior dev pushes to prod on Friday because they know the on-call engineer owes them a beer
Treating prod like a Spanner table - eventually consistent with your weekend plans
The same confidence that led me to skip canary deployments is what's letting me ignore seventeen PagerDuty alerts on my phone right now
Friday 5pm deploys are fine - the rollback plan is the on-call engineer's weekend, and that's technically documented in the runbook as 'TBD'
The Friday 5pm production deployment: where 'move fast and break things' meets 'I'll be unreachable until Monday.' It's the perfect storm of optimism bias and weekend plans - because nothing says 'I trust my CI/CD pipeline' quite like deploying a major release right before you turn off Slack notifications. Bonus points if you're also doing a database migration and the rollback plan is 'restore from backup... probably.'
Shipping a Friday 16:59 schema migration with alerts muted and no canary is basically converting your error budget into a majestic plume - rollback now requires a data backfill
Friday 5PM prod deploy: the release strategy where 'eventual consistency' means Monday morning triage
Calling it a ‘canary’ because only half the team is on call isn’t what SRE meant by reducing blast radius