From e00012be3dc50bd54d5146732e660d49cd781b72 Mon Sep 17 00:00:00 2001 From: "molecule-ai[bot]" <276602405+molecule-ai[bot]@users.noreply.github.com> Date: Tue, 21 Apr 2026 03:38:29 +0000 Subject: [PATCH] docs(blog): debug failed workspaces without leaving Canvas MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Content for internal#8 / PR #1178 launch coverage. Angle: EC2 console output surfaced directly in Canvas — no more tab-switching to AWS Console. ~180 words. --- .../index.mdx | 20 +++++++++++++++++++ 1 file changed, 20 insertions(+) create mode 100644 content/blog/2026-04-21-failed-workspace-ec2-console/index.mdx diff --git a/content/blog/2026-04-21-failed-workspace-ec2-console/index.mdx b/content/blog/2026-04-21-failed-workspace-ec2-console/index.mdx new file mode 100644 index 0000000..70af4e9 --- /dev/null +++ b/content/blog/2026-04-21-failed-workspace-ec2-console/index.mdx @@ -0,0 +1,20 @@ +--- +title: "Debug Failed Workspaces Without Leaving Canvas" +description: "Molecule AI Canvas now surfaces EC2 console output directly in the failed workspace view — no more tabbing to the AWS Console to understand what went wrong." +publishedAt: 2026-04-21 +tags: [canvas, devops, observability, ec2, workspaces] +--- + +When a Molecule AI workspace fails, the reason lives in the AWS EC2 console. Until today, that meant opening the AWS Console, finding the right instance, pulling the console log, and figuring out what happened — while your workspace sat failed and your team waited. + +That's now unnecessary. + +Canvas now surfaces `last_sample_error` and the full EC2 console output directly in the failed workspace view. You open the error panel, read the output, identify the root cause, and fix it — all without leaving Molecule AI. + +The trigger is a workspace that transitions to `failed` state. Molecule AI captures the EC2 console log from that point and attaches it to the workspace record. It appears in the error panel as soon as the failure is detected, ready to read. + +The impact is simple: one tool, one context. Debug in Canvas, fix, redeploy. + +If you're self-hosting Molecule AI, update to the latest release and the feature is active immediately. For hosted deployments, it's live now. + +Next time a workspace fails, open the error panel before you open AWS. You'll find what you need there.