The dashboard builds itself
Deploy a workflow and its metrics are already there. Throughput, errors and latency roll up per workspace and break down per workflow — and the breakdown names the node responsible for the most errors, so root cause is on the first screen.
Throughput, errors and latency per workspace and per workflow, with the failing node named for you and no dashboard to build.
What you can see
- A dashboard you did not have to build — Summary counters and a time series arrive with the workspace, over any range you pick. There is no metric to instrument, no panel to lay out and no query language to learn before you can answer whether last night went well.
- Root cause without the drilldown — For any workflow with errors in range, the per-workflow breakdown also reports the node that produced the most of them, and how many. The usual loop — open the workflow, open a failed run, hunt for the step — collapses into one line on the overview.
- Live in-flight count — See how many executions are running right now, not just how many finished. It is the difference between knowing a queue is draining and finding out tomorrow that it never was.
- Run-level inspection — Open any execution and walk it node by node, with the payload that entered and left each step. The same view is available in the free sandbox before you sign up, so you can see what debugging actually looks like first.
Platform capabilities
- 0: Dashboards you have to configure
- Per-node: Error attribution on the overview
- Live: Count of executions in flight
Do I have to set up monitoring before I can use it?
No. Metrics are collected from executions as they run, and the workspace summary, the time series and the per-workflow breakdown are all present from the first run. There is nothing to instrument and no dashboard to assemble.
How do I find which step is causing the failures?
The overview tells you. For every workflow with errors in the selected range, the breakdown includes the node that accounted for the most of them along with its error count, so you start from the answer instead of searching for it.
Can I try the monitoring view before signing up?
Yes. The sandbox runs real webhook traffic through a real graph and shows the same run inspection view, with no account and no credit card. It is the fastest way to judge whether the debugging experience is what you want.