Monitoring Automations
Monitoring helps you answer one simple question:
"What is happening with my automation right now?"
After an automation starts, you can watch its progress, check each step, see the final result, and understand what went wrong if it fails.
The Basic Idea
What You Can See
| Area | What it tells you |
|---|---|
| Overall status | Whether the automation is waiting, running, completed, failed, paused, or stopped |
| Current step | Which part of the automation is running now |
| Step results | What each step produced after it finished |
| Duration | How long the run or step took |
| Errors | What went wrong and where it happened |
| Recent runs | Whether the automation usually succeeds or often fails |
The Executions View
Every job has an Executions tab (next to Configurations and Canvas on the job canvas). This is where you watch runs live and inspect past ones.
Auto-refresh
The Executions panel can refresh itself on an interval. Click the refresh badge to open the Auto-refresh menu and pick how often the list updates:
| Option | Behavior |
|---|---|
Off | The list only updates when you refresh it manually |
Every 5s / 10s / 30s | Frequent updates for a run you're actively watching |
Every 1m / 2m / 5m | Lighter polling for background monitoring |
The badge shows the current interval, and a ring animates around the refresh button while auto-refresh is active. You can also refresh once at any time with the refresh button.
Stat Cards
A compact stats card sits at the top of the Executions panel and summarizes the job's recent health:
| Card | What it shows |
|---|---|
| Last Run | The status of the most recent run and how long ago it happened |
| Fail Rate | The share of recent runs that failed |
| Avg Dur | The average run duration, in seconds |
Filtering Runs
Click the funnel icon to open the filter panel. You can narrow the list by:
| Filter | Options |
|---|---|
| Status | All, COMPLETED, RUNNING, FAILED |
| Time range | Today, Last (a number plus a unit), or Custom (a From/To range) |
| Search executions | Free-text match on ID, user, payload, or error text |
| Duration | An operator, a value, and a unit (e.g. longer than 30 seconds) |
Use Apply to run the filter, Reset to clear it, or Cancel to close the panel. Each run in the list shows its trigger type, start time, status, and duration in seconds.
Inspecting a Run
Selecting a run opens its details. You can view a run in Inspect or Flow mode, switch the timezone used for timestamps, and open the Output, Config, and Logs tabs for any step - each viewable as formatted JSON or as a table.
Timeline
Toggle Timeline to see the run laid out as a horizontal bar chart. It shows the total run duration and one bar per step, positioned by when the step started and sized by how long it took, with each bar colored by the step's status. This makes it easy to spot the slowest steps and which steps overlapped (ran in parallel).
Per-run Controls
Each run in the list has action buttons that change based on its current status:
| Run status | Available actions |
|---|---|
RUNNING | Pause (hold the run) or Stop (end it) |
PAUSED | Resume (continue) or Stop (end it) |
FAILED | Restart (run it again) |
COMPLETED | Retry (run it again) |
Status Cheat Sheet
| Status | Meaning | What to do |
|---|---|---|
PENDING | The automation is waiting to start | Check again shortly |
RUNNING | The automation is working now | Watch progress |
COMPLETED | All steps finished successfully | Review the result |
FAILED | One or more steps did not finish | Open the failed step and read the error |
PAUSED | The run is waiting before continuing | Resume it when ready |
STOPPED | The run was stopped manually | Review what completed before it stopped |
Example Run
Here is how a successful report automation may look:
The full automation completed in 11 seconds.
How to Read Step Results
Each step can produce information that later steps use.
| Step | Output | Why it matters |
|---|---|---|
| Query data | Customer, sales, or ticket records | Later steps use this data |
| Create file | Report, spreadsheet, or attachment | A later step may send or store it |
| Send message | Delivery result | Confirms who received the message |
| Update record | Updated status or ID | Confirms the business change happened |
If a later step fails, always check the output from the step before it. Many failures happen because the previous step returned missing or unexpected data.
When Something Fails
Start with these questions:
| Question | Why it helps |
|---|---|
| Which step failed first? | The first failed step usually shows the real problem |
| What does the error say? | The message often tells you exactly what was missing or unavailable |
| Did the previous step produce the expected output? | Bad or missing input can break the next step |
| Did it fail quickly or after waiting? | Quick failures often mean setup issues; slow failures often mean timeout or connection issues |
| Did this start after a recent change? | A workflow, password, file path, or external system may have changed |
Common Things to Check
| What you see | Likely reason | Next step |
|---|---|---|
| Connection timeout | External system was slow or unreachable | Try again or check system availability |
| Authentication failed | Credentials or permission may be wrong | Review the connection setup |
| No data returned | The query worked but found nothing | Check date, filter, or source data |
| File not found | A file path is wrong or a previous step did not create it | Check the previous step output |
| Same step fails every time | Configuration or input is likely wrong | Review that step before retrying |
Useful Monitoring Questions
Did today's job succeed?
Check the latest run and confirm the status is COMPLETED.
Why did this automation fail?
Open the failed run, find the first failed step, and read the error message plus the previous step output.
Is this automation getting slower?
Compare recent run durations. If the same step is taking longer each time, that step may need attention.
Is the schedule working?
Review recent run times and confirm they match the expected schedule.
What to Watch Over Time
| Metric | Healthy sign | Warning sign |
|---|---|---|
| Success rate | Most runs complete successfully | Failures are increasing |
| Duration | Runs take a predictable amount of time | Runs are getting slower |
| Failed step | Failures happen rarely and in different places | The same step fails repeatedly |
| Output quality | Results look complete and expected | Data is missing, duplicated, or unexpected |
| Queue time | Runs start quickly | Runs wait too long before starting |
Good Monitoring Habits
Check the first failed step
Do not start at the last error. The first failed step is usually the most useful clue.
Look at output, not only status
A run can complete but still produce unexpected data. Review important outputs.
Compare with a successful run
If something changed, a previous successful run can show what the normal pattern looked like.
Watch repeated problems
One failure may be temporary. The same failure again and again usually means the automation needs a fix.
Related Pages
- Retry Strategies - how automations can try again after temporary failures
- Concurrency Control - how to prevent too many runs at the same time
- Logging - how to read detailed step messages