Shared signal for trusted workflows
Glubean Cloud does not create a second system. It takes the same workflows you already run in VS Code and CI, then adds shared history, pass-rate trends, latency tracking, alerts, and status surfaces the team can actually use. Secrets are redacted before upload, so the signal stays shareable.
Cloud overview
same workflow, now visible to the team
Latest workflow runs
synced across surfacesTeam outcomes
+4.2% this week
-35% after fix
custom metric from header
email + webhook
VS Code
Author and debug
CI
Gate the same file
Cloud
Keep the signal shared
Why Cloud exists
Teams usually trust the workflow itself before they trust the reporting around it. Cloud should feel like a continuation of that workflow, not a dashboard you have to learn from scratch.
Without Cloud
Local runs disappear after you close the editor.
CI logs tell you a workflow failed, but not whether reliability is drifting.
Teams end up rebuilding history, alerts, and dashboards around the same workflow.
With Cloud
The same workflow accumulates history instead of disappearing into logs.
Pass rate and latency tell the team whether reliability is improving or drifting.
Alerts, public status, and badges reuse the workflow the team already understands.
One workflow, three surfaces
That is the real job of Cloud. The workflow starts in one place, gets gated in another, and becomes shared operational signal for the team without turning into a different artifact.
VS Code
Write or refine the workflow in the editor and inspect traces while the context is still fresh.
CI
Run the identical workflow in pull requests and scheduled jobs with the CLI. No rewrite, no export step.
Cloud
Keep history, pass rate, latency, and notifications attached to the workflow the team already trusts.
What teams actually get
Each capability should answer a team question: what broke, is it getting worse, who needs to know, and how do we share the answer?
Track every workflow over time instead of treating each CI run like an isolated event.
See whether a workflow is still healthy, getting slower, or starting to drift before users complain.
Track anything from response headers to algorithm scores with ctx.metric(). Cloud aggregates trends, detects regressions, and alerts when thresholds break.
Email and webhook notifications stay attached to the exact workflow, run, and trace that triggered them.
Expose badges and public dashboards without creating a second reporting system for the team.
Active incident
refund-flow
Failure started after deploy `2f4e1ad`. P95 climbed from 184ms to 611ms.
Notification fanout
Public status
Prometheus export
/metrics/glubean/p95?workflow=checkout-flow
Open API
/open/v1/projects/:id/runs
Security
Glubean separates credentials from configuration at the file level and redacts sensitive values before anything is uploaded. Cloud receives test results — assertions, traces, metrics — but never your API keys, tokens, or passwords.
Learn how redaction worksFile-level separation
.env
BASE_URL, REGION, flags
.env.secrets
API_KEY, TOKEN — gitignored
ctx.vars vs ctx.secrets — the runner knows which values are sensitive.
Custom redaction engine
Authorization: Bearer [REDACTED]
api_key: [REDACTED:key]
token: sk_li***_4xN
Dual-layer detection: sensitive keys + value patterns (JWT, AWS, GitHub PAT).
22 built-in rules that cannot be weakened — only extended.
Preview with glubean redact before uploading.