tl;dr - cloud infrastructure is no longer whack-a-mole but closer to tower defense. Today's teams care more about proactive fixes (platform engineering) than incident response (SRE). @flightcrew.io helps them do that

SREs are becoming SLO coaches and managing log pipelines ... much less firefighting!

Staffing Ratios are changing to reflect this. We’re seeing 1 SRE for every 10 Platform Engineers, for every 100 Feature Engineers.

Proposed ... #platformengineering has significantly changed the role (and value) of SREs

1️⃣ Most incidents are triggered by rollouts
2️⃣ Platform Engineering have made rollouts much safer (via B/G deploys, dev environments, etc)
3️⃣ Incidents are less frequent, and more easily classified

Eyed Wide Shut is borderline

Lethal Weapon is not

What if we’re not enthusiasts, but cynical AWS drones. Still okay?

We got on Bluesky bc it's impossible to claim your trademarks/copyrights as handles on X post-Elon. Regardless of your stance on national socialism, Bluesky is a better place to build a business

PS @ediri.de had a good post last week on using raw yaml vs TF/OT/HCL vs Pulumi www.pulumi.com/blog/yaml-te...

IF YOU ARE JOINING AN EARLY STAGE STARTUP FOR INFRA - always ask how they're actually deploying Kubernetes/infra. Like ask 5 times until you actually know. You do not want to be responsible for maintaining or fixing custom systems

You want open source! And something that all of your engineers can understand and modify without DMing you. If you like abstractions try Pulumi, Dagger, etc ... just don't build your own!

Custom deployment systems are brittle, expensive to maintain, and incredibly difficult for your broader engineering team to tweak and extend. It's a huge knowledge/maintenance problem that kills platform teams