Metrics
See how every workflow in a workspace ran over a time range, and where failures and spend concentrate
The Metrics page shows how every workflow in a workspace ran over a chosen time range. It reports volume, reliability, latency, and cost per workflow, and points at the workflows where failures and wasted spend sit.
Metrics summarizes runs. To inspect one run, open Logs. To see credits and billing, open Settings → Subscription and read Cost Calculation.
Opening Metrics
Click Metrics in the workspace sidebar. Every workspace member with read access can open the page. The URL is /{orgSlug}/{workspaceSlug}/metrics.
Time Range
The selector at the top right sets the range for the whole page. The default is Past 7 days. The other options are:
- All time
- Past 30 minutes, Past hour, Past 6 hours, Past 12 hours, Past 24 hours
- Past 3 days, Past 14 days, Past 30 days
The range is stored in the URL, so a copied link opens the same view.
The page has no custom date range. For an exact window, use the date filter on the Logs page.
What the Page Shows
Headline figures: four numbers for the range. Runs is the total number of executions. Succeeded is the success rate, with failed, cancelled, running, and pending counts under it. p90 latency is the time within which 90% of timed runs finished. Cost is the total execution cost, with the wasted spend from failed runs under it.
Story line: one sentence under the figures that names the workflows with the most failures and wasted spend. Hover a name to highlight that workflow in the chart. Click a name to expand its row.
Run chart: runs over time, stacked by workflow. Ticks on the chart mark deployments inside the range, labelled with the version number such as v12.
Trigger mix: a strip that shows the share of runs per trigger type, for example API, Webhook, Schedule, Manual, or Chat.
Workflow table: one row per workflow with Runs, Success, p90, Cost, Wasted, and Last run. Click a column header to sort. Workflows with no run in the range are listed at the end. Their row states the reason, such as "No run in past 7 days" or "Never ran".
Expand a row to see its Outcomes, Latency, and Cost compared with the workspace as a whole. When the latest deployment has enough runs, the row also compares it with the previous version. Each row has links to Open workflow and to its failed runs in Logs.
How Figures Are Counted
- Runs counts every execution of the workspace's active workflows that started in the range, whatever the outcome. Runs of archived or deleted workflows are not included, so totals can differ from the Logs page.
- Success, failure, and cancelled rates are percentages of finished runs. A finished run completed, failed, or was cancelled. Runs that are still running or waiting for approval do not count against reliability.
- Latency and Cost come only from runs that carry timing data. When no such runs exist, the page shows "No timing samples in range".
- Cost is the duration-based execution cost of timed runs. Wasted is the same cost counted only for failed runs. These are not your credit balance.
Comparing Workspaces
Click Compare workspaces on the workspace Metrics page to open organization Metrics at /{orgSlug}/metrics. Any organization member can open it. The page lists only the workspaces you can access.
Select the workspaces to compare. The page shows one row per workspace and one rolled-up total and chart for the selection. You can compare up to 20 workspaces at a time. A larger selection shows "Select fewer workspaces to compare" instead of a partial result. The same time range options apply.