Workflow Execution Observability
A real-time dashboard and execution history for every workflow run. See what's running, what's waiting on approval, and what failed — with step-level logs, durations, and failure reasons inline.
Once you start running automations against production, "does it work?" isn't enough. You need to know what ran, what's running right now, what's waiting on a human, what failed and why — and which workflows are doing the heavy lifting versus quietly breaking. Today we're shipping execution observability for Kestrel Workflows: a real-time dashboard, a full execution history, and step-level detail for every run.
A bird's-eye view across every workflow
The Observability dashboard opens with the numbers that matter at a glance: Executions (30d), Success Rate, Avg Duration, Avg Approval Wait, and Active Workflows.
Below that, a set of charts breaks down what's actually happening across your organization — execution volume over time by workflow, a ranking of your most-used workflows, failure rate by workflow, execution duration, integration usage, trigger sources, approval wait times, and approval outcomes. You can see at a glance which workflows run the most and which ones need attention.
Every run, in one list
The Executions page lists every workflow run across your organization, so you always know exactly what's running and what's waiting for approval. Each row shows the Workflow, its Status, what triggered it, when it started, and how long it took. If a run failed, the error is shown inline — you don't have to click in to find out what went wrong.
Filter the list by status (Completed, Failed, Running, Awaiting Approval, Cancelled), by workflow name, or by trigger source — Kubernetes, AWS, PagerDuty, PostHog, Vercel, Slack requests, and more — or search across workflow names and error messages.
Step-by-step, in real time
Click into any execution and you see the workflow as a DAG, with real-time status on each step. Steps show their duration and a green checkmark as they complete, the current step is highlighted, and any failure is surfaced right on the node.
Click any block to inspect it. For the trigger, you see the original signal — what happened, the affected resource, severity, and the exact timestamp. For action steps, you see the full output: AI investigation results and root cause, generated fixes, API responses, and any resources that were created. You get complete visibility into everything your workflow did, step by step.
When something fails
Failures are a first-class part of the view, not an afterthought. The Executions list shows the error inline, the failed step is marked on the DAG, and the step inspector shows the exact error. From a failed run you can replay it — from the beginning or from the failed step — and open the failure diagnosis agent to investigate what went wrong and generate a corrected workflow you can save with one click. (More on that agent in an upcoming changelog.)
Getting Started
Observability is on by default for every Kestrel Workflows account. Open Workflows → Observability for the dashboard, or Workflows → Executions to browse runs. Every execution — past, live, or waiting — is already there.