Skip to content
Docs
flero.ai

Log explorer

/logs is the global execution search across every workflow in your workspace. Inspired by BigQuery / Datadog: query builder, saved queries, status chips, date picker, expandable rows.

📸 Screenshot needed: execution__log-explorer.png, /logs screen with the query builder open on the right (one filter line: status = failed, another: workflow contains "Slack"), status filter chips highlighted at top, date range "Last 7 days", and a table of 12 rows below.

Layout

┌──────────────────────────────────────────────────────────────────────┐
│ Toolbar  [Search] [Sort] [Columns] [Date range] [Saved queries] [⟳] │
├──────────────────────────────────────────────────────────┬───────────┤
│ Status chips: All  Success  Failed  Running  Cancelled   │ Query     │
├──────────────────────────────────────────────────────────┤ builder   │
│ Results table                                            │           │
│                                                          │           │
│ ID   Workflow   Status   Started   Duration   Trigger…  │           │
│                                                          │           │
└──────────────────────────────────────────────────────────┴───────────┘

Toolbar

  • Search, full-text search across execution ID, workflow name, trigger payload, and error messages.
  • Sort, by date (default desc), duration, status, workflow.
  • Columns, choose which columns are visible.
  • Date range, last 1 h / 24 h / 7 d / 30 d / 90 d / custom.
  • Saved queries, see Saved queries below.
  • Refresh ⟳, manual; toggle Live for 3-second polling.
  • Export, CSV / JSON / NDJSON.

Status chips

Quick filters across the top: All | Success | Failed | Running | Cancelled | Timed out | Waiting

Click one to narrow. Combine with the query builder for finer cuts.

Query builder (right panel)

A stack of filter rows. Each row picks a field, an operator, and a value:

Field Example operators Example values
status equals, not equals success, failed, …
workflow.name equals, contains, regex Hourly sync
workflow.tags contains production
trigger.type equals webhook, schedule, …
duration >, <, between 1000 ms
error.type equals, contains TimeoutError
cost.tokens > 10000
started_at between datetime range
node.<name>.status equals failed

Combine rows with AND / OR (click the toggle between rows). Nest with the ( ) button.

Results table

Columns (configurable):

  • Execution ID
  • Workflow
  • Status
  • Started
  • Duration
  • Trigger
  • Last node
  • Output summary
  • Cost (if AI nodes ran)

Click a row to expand inline (compact view of the timeline). Click again or use the row's Open action to navigate to the full execution detail page.

Saved queries

Save a query for reuse:

  1. Build your filter set.
  2. Click Save query at the top of the query builder.
  3. Name it ("Failed prod runs last 24 h", "High-cost AI runs").
  4. Choose share scope (Personal / Workspace).

Saved queries appear in the dropdown at the top of the toolbar. One-click recall.

Useful saved queries to set up early:

  • Failed in the last 24 hours, status = failed AND started_at > now() - 1d
  • Long-running, duration > 60000 AND status != running
  • High AI cost, cost.tokens > 50000
  • Webhooks from {service}, trigger.type = webhook AND trigger.path contains "{service}"

Schema browser

Click the Schema button in the query builder to see every field available, including custom fields you've added via metadata. Useful when you forget the exact field name.

Exports

  • CSV, flat columns; good for spreadsheets.
  • JSON, full structured records; good for downstream tooling.
  • NDJSON, one record per line; good for piping into another system.

Exports respect the current filter. Maximum 10 000 rows per export, refine your query if you need more.

Tips & gotchas

  • Cost field is empty for workflows without AI nodes. Filter cost.tokens > 0 to focus on AI runs.
  • node.<name>.status filters require the exact node name as it was at execution time. Renames don't retroactively rename in history.
  • Live refresh in a busy workspace can be expensive. Use targeted filters before turning it on.
  • The explorer is read-only. To actually fix something, drill into a failed run and use the Replay action.

Found something out of date? This page lives in the Flero docs content set.