Solving Meta's top 4 outage causes: 1/4 Unexpected Dependencies

In 2022 I watched Francois Richard deliver an excellent talk at SRECon EMEA about how Meta drained every backbone router simultaneously (you remember, it was *that* outage). Here were their top 4 trending root causes of outages:

At Overmind we've had our heads down building and getting feedback over the past few months and I think we're pretty close to solving one of the four: Unexpected Dependencies

Here's an update of what we've built since I last posted an update.

How? Blast Radius.

Use our?GitHub action (or do it manually)?to go from?Terraform Plan --> Blast Radius. The blast radius is based on your live AWS state, not Terraform, which lets you see what might break:

• Includes?resources not managed by Terraform
• Discovers dependencies?even if they were created manually
• Shows?live data, not out-of-date CMDB data
• Does all of this with read-only access, no agents, no telemetry, and no input from you. If you had to tell us how your apps are architected, we're hardly going to find unexpected dependencies are we?

What's next? (2/4) Configuration Updates

We're not stopping with just blast radius. Once you've decided to apply your changes, track them with Overmind. Since we've already worked out all the dependencies, we can tell you if your changes has broken something downstream, even if you didn't know it existed.

Want to try it?

Sign up and start calculating blast radius now! We’d love any feedback. Either by email, discord, or?book a meeting.

Note: This is beta software & we’d love any feedback. Either by Discord, or?book a meeting. It's free for individuals, if you're interested in a team plan contact me.