Monitoring

Real-time execution logs, status panels, and error tracking across all running workflows.

Overview

The Monitoring screen gives you a live view of every workflow execution happening across TotalApp. Execution events stream in real time — you can watch each pipeline step transition from pending to running to done without refreshing the page.

Use Monitoring to catch failures immediately, review the output of recent runs, and understand the overall health of your automation estate.

Execution Log

The central log panel lists every execution event in reverse-chronological order. Each entry shows:

Workflow name — which pipeline generated this event.
Node name — which step within the pipeline.
Status — pending, running, done, or error.
Timestamp — when the event occurred.
Duration — how long the node took to complete.
Output preview — a truncated view of the node's output value (click to expand).

Filter by Status

Use the status filter pills at the top of the log to show only errors, only running steps, or only completed steps. This is the fastest way to triage a failing pipeline.

Status Panel

The status panel on the right summarises key metrics across all workflows:

Active runs — pipelines currently executing.
Completed (last hour) — successful executions in the past 60 minutes.
Errors (last hour) — failed steps in the past 60 minutes.
Average duration — mean execution time per workflow run.

Error Details

When a node fails, the monitoring log highlights the entry in red. Click the entry to expand the full error message, including the stack trace returned by the node executor. Common causes include:

Missing API key for an AI node (check Settings → API keys).
Network timeout from an external API Source node.
JavaScript runtime error inside a Script node.
Empty input passed to a node that requires non-empty input.

Downstream Nodes Are Skipped on Error

When a node errors, all downstream nodes that depend on its output are automatically skipped. The workflow marks those nodes as cancelled, not errored. Only the originating failure shows an error status in the log.

AI Assistant

Monitoring ships with a built-in AI Assistant. Click the vertical AI Assistant tab on the right edge of the screen (or the Sparkles toggle) to open a chat panel that reads your current execution stats and recent runs. Because Monitoring is a dashboard/overview screen rather than a record-editing table, the assistant is scoped to dashboard-style capabilities only — it narrates and explains the numbers already shown, it never edits workflows or executions:

Executive summary — a plain-language readout of the success rate and the most notable recent failures or warnings, by workflow name.
Root cause & anomaly detection — when failures or warnings are elevated, the assistant explains which workflows are responsible and whether there is a pattern (e.g. the same workflow repeatedly failing, or slow durations).
Recommendations — prioritizes which failing or warned workflows to investigate first, referencing the actual workflow names and statuses.

Every reply can be saved with the Save as Report button, which stores it in My Reports under a title derived from your question and the current month/year. The panel respects the Writer Engine selected in Settings → Agentic (Local CLI, Hosted API, Local LLM, or Ollama) — Ollama and other local/in-browser engines run entirely on your own machine and never proxy through the server.

Frequently Asked Questions

Does the execution log update automatically, or do I need to refresh the page?

The log streams in real time — new events for pending, running, done, and error states appear as they happen without a manual page refresh. Keep the Monitoring screen open in a tab while a critical workflow runs to watch its progress live.

Why did my workflow show a mix of "error" and "cancelled" statuses instead of all "error"?

Only the node that actually failed is marked as error. Every downstream node that depended on its output is automatically skipped and marked cancelled rather than error, since it never had valid input to execute with. Look for the single error entry to find the true root cause — cancelled entries are a side effect, not separate failures.

How far back does the execution log retain history?

The Status Panel summarises Active runs, Completed, and Errors for the last hour by default, but the log itself lists all recent execution events in reverse-chronological order. Use the status filter pills to narrow a long history down to just errors or just running steps when triaging.

I see an error but the message is cut off — how do I see the full stack trace?

Click the log entry to expand it. The expanded view shows the complete error message and stack trace returned by the node executor, which is usually enough to identify the cause — a missing API key, a network timeout on an API Source node, a Script node runtime error, or empty input passed to a node that requires a value.