Edera

Edera

@edera.dev

We just want to talk about hardened runtime and stuff. GitHub: github.com/edera-dev YouTube: youtube.com/@edera-dev Website: edera.dev

311 followers74 following301 posts

View full thread

“At the point we are freezing and CoW, we are talking milliseconds, whereas Kubernetes frequently measures things in seconds.” Getting there took custom tooling for freezing and cloning memory state. Alex: “We are often optimizing Kubernetes itself to make it all possible.”

Live snapshotting lets an agent waiting on human input save its full state, suspend, and resume when needed. Teams can also prepare an environment once and reuse it, instead of repeating resource-intensive setup for every task or RL run.

Copy-on-write forking clones a running microVM into isolated copies that share memory until they change a page. Each copy is a zone, created with one hypervisor call. “We can fork thousands of zones very quickly with this design,” says Edera CTO @alex.zenla.io

48% of organizations using AI coding assistants say more security issues are reaching production since they adopted them. Detection tools were trained on human-written code. They have no baseline for code written by agents. edera.link/idc-brief

edera.linkNew IDC Analyst Brief: Containment for AI Workloads | EderaPrevention can't keep pace with AI-written code. A new IDC Analyst Brief explains why containment—bounding a breach's blast radius—is now the control for AI workloads.

As agents write and run more code, more of it goes unexamined. IDC study's take: “Containment doesn't ask enterprises to verify it first; it bounds what any of it can affect, using infrastructure most enterprises have already deployed at scale.” . Read the analyst brief from IDC

edera.linkNew IDC Analyst Brief: Containment for AI Workloads | EderaPrevention can't keep pace with AI-written code. A new IDC Analyst Brief explains why containment—bounding a breach's blast radius—is now the control for AI workloads.

"Sandbox" has become one of the most overloaded words in Kubernetes, and now it feels like it can mean any of a dozen distinct things. On October 29, we're hosting a live panel on the isolation landscape moderated by @alex.zenla.io. edera.link/vev-1029

72% of enterprises are already running AI agents in production. Most of the controls protecting them assume a compromise can be stopped before it happens. That assumption breaks once systems and agents act on their own. edera.link/idc-brief

An AI agent turned one exploit into a full production breach over a weekend. Shared-kernel isolation didn't hold. Alex Zenla and Claude Mandy get into it live, Sep 24 at 12pm ET edera.link/vev-923

Edera reposted

I had a lot of fun building some cool, futuristic stuff with Marina to achieve builds with attested hermeticity.

Edera@edera.dev

Did you know we have a YouTube channel where we post all our webinars to watch on-demand without any obnoxious forms to fill out? Anyway, here's the latest one we did with special guest @puerco.mx ! edera.link/vev-820

Edera reposted

Replying to @whitneylee.com

I love SPIFFE/SPIRE! Last week I did a cool demo with Marina Moore from @edera.dev where we ran hermetic builds attested and signed with SPIFFE SVIDs, check it out! bsky.app/profile/eder...

Edera@edera.dev

SLSA says builds should run isolated. Can you prove yours did? Aug 19, 11am ET: Marina Moore & @puerco.mx show how to verify hermetic builds: cryptographic proof, full SBOM coverage, provenance to the exact VM that ran it. Powered by Cocoon in Edera Protect zones. edera.link/vev-820

SLSA says builds should run isolated. Can you prove yours did? Aug 19, 11am ET: Marina Moore & @puerco.mx show how to verify hermetic builds: cryptographic proof, full SBOM coverage, provenance to the exact VM that ran it. Powered by Cocoon in Edera Protect zones. edera.link/vev-820

edera.linkBuilt Clean. Receipts Attached.Security frameworks such as SLSA require software builds to run in isolated environments to guarantee they are “free of unintended external influence”. Except....

Edera reposted

Join us on Wednesday, August 5, 7:00–10:00 PM at for the @carahsoft.bsky.social Public Sector Reception at Minus5° ICEBAR. Cocktails, conversation, and a break from the Vegas heat inside 90 tons of ice. Space is limited. #blackhat Register Now: edera.link/i8dpYqw

Edera reposted

Bob Lord (CISA Secure by Design co-lead) joins our CTO @alex.zenla.io and Corridor's CTO @ashwinramaswami.bsky.social to talk about securing AI-era software from build time to runtime. Moderated by IDC's @katienorton.bsky.social Come for the panel, stay for the drinks. #BlackHat

edera.linkCorridor and Edera Party · LumaCorridor and Edera are teaming up to spill the tea on the hardest problem in AI software: how do you trust code that machines wrote? AI writes code faster than…

Bob Lord (CISA Secure by Design co-lead) joins our CTO @alex.zenla.io and Corridor's CTO @ashwinramaswami.bsky.social to talk about securing AI-era software from build time to runtime. Moderated by IDC's @katienorton.bsky.social Come for the panel, stay for the drinks. #BlackHat

edera.linkCorridor and Edera Party · LumaCorridor and Edera are teaming up to spill the tea on the hardest problem in AI software: how do you trust code that machines wrote? AI writes code faster than…

Standard Kubernetes metrics stop at the container boundary. Kernel pressure, page cache eviction, I/O stalls — all invisible per-workload on a shared kernel. @jsalazar.co breaks down kernel-per-workload observability. Aug 6, 1pm ET edera.link/vev-806

edera.linkKernel Observability: The Missing Layer in Cloud Native EngineeringPlatform engineering has transformed how we deliver Kubernetes to development teams. We've abstracted away node management, networking complexity, and deployment pipelines. But we've accidentally crea...

Edera reposted

Have you ever said to yourself "Dang, that axolotl at Edera is pretty cool, but I really just want to test out the tool"? Well have we got news for you....

Have you ever said to yourself "Dang, that axolotl at Edera is pretty cool, but I really just want to test out the tool"? Well have we got news for you....

View full thread

Have you ever said to yourself "Dang, that axolotl at Edera is pretty cool, but I really just want to test out the tool"? Well have we got news for you....

If you're PlatformCon NYC today, be sure to join us this evening. If you're a footy fanatic, we'll have the matches on. If you couldn't care less, we'll have good chat, snacks, and drinks on us! luma.com/i7f1o8h2

luma.comWatch Party & Happy Hour · LumaWhen PlatformCon ends, the World Cup is still going. Join Edera, Minimus, and Dash0 for drinks, live match coverage, and good company at DB Coopers in Hudson…
Edera@edera.dev

The matches are on. The drinks are on us. Edera, @minimus.io & @dash0.com are hosting a World Cup happy hour at DB Coopers during PlatformCon NYC. Space is limited, RSVP luma.com/i7f1o8h2

Edera reposted

This Sunday, @edera.dev is hosting a luxury suite at the Valkyries-Liberty WNBA game in SF and we have a few extra tickets left! If you work in cybersecurity or the general tech space, DM me if you might be interested! I'll be there as well!

GPUs weren't built for multi-tenancy. Marina Moore at PlatformCon covers what that actually means: memory leakage, hardware vs. software isolation, vGPUs, and vendor lock-in's security implications. Worth your time if you run shared AI infra. run.edera.dev/platformcon-...

Edera reposted

having watched VMs go through the cycle of getting efficient and then getting protected even from admins, I've always been a bit salty about whether containers had enough isolation and it's been interesting to see projects emerge that suggest people agree with me: now AI agents are forcing the issue

thestack.technologyHow to get visibility and isolation for AI in KubernetesAI is speed running virtualisation and cloud computing, but how do you isolate for security and still get enough information to make good decisions?

Edera reposted

Replying to @marypcbuk.bsky.social

GPUs can use the new Dynamic Resource Allocation API to tell Kubernetes about themselves so you can do more than, say, allocate a whole GPU; they can also report back metrics. @alex.zenla.io tells me how @edera.dev uses DRA to get all the metrics you want out of containers in a Kubernetes native way

thestack.technologyHow to get visibility and isolation for AI in KubernetesAI is speed running virtualisation and cloud computing, but how do you isolate for security and still get enough information to make good decisions?