What this solves#
A collection runs every night at 2am. One night, yesterday's run is still going when 2am arrives.
- Overlapping them pulls the same files twice
- Skipping today leaves today's data empty
- Letting it pass unnoticed means finding out in the morning
All three are handled differently. And beyond overlap, the product records "the schedule never fired", "it started late" and "the previous stage did not finish" under separate names.
This document covers setting a schedule, what happens when runs collide, and where to look.
Who this is for#
| Role | When |
|---|---|
| Operations | Running daily or weekly transfers |
| System administrator | Night work sometimes slips into the morning |
| Business owner | Data has to be there by a fixed time |
Setting the schedule#
Flows → pick the Flow → Edit flow.
Under Configuration on the right, choose Date/time in Start when.
Enter the date and time in Start time.
Pick the cadence in Repeat.
Option Meaning NeverOnce only Daily at <time>Every day at that time Weekly on <day>Every week on that day Monthly on the <date>Every month on that date Every weekday (Monday to Friday)Weekdays only Every weekend (Saturday and Sunday)Weekends only CustomYour own interval The weekday and date come from the date you entered in Start time.
Running more than once a day is only under
Custom. The list above has no hourly entry. PickCustomand a second selector appears withHourly,Daily,WeeklyandMonthly, followed by Every and a number (1 to 99), so "every 4 hours" isCustom→Hourly→Every 4.This list is a Flow's
Repeat. If you need no branching and no extra steps, only a repeating transfer, there is a second path: start a transfer from Devices (the window is titledTransfer) and its schedule step offersOne time only,Every hour,Every day,Every weekandEvery month— hourly is right there in the list. Everything below about overlaps and missed runs applies to both paths.Pick the Time zone. Below the field the chosen zone and the note run together on one line - "America/New_York · Automatically adjusts for daylight saving time" (no full stop). The field shows just the city, the line below it the full region name.
Press Save & activate.
With sites in several countries, set the time zone to the one the business runs on. Scheduling in head-office time means the transfer starts during opening hours locally.
What happens when runs collide#
The Schedule and time group sits under Error handling in Assurance. Below are INNORIX's recommended values.
| Issue | Default | Meaning |
|---|---|---|
Schedule overlap |
Skip this run |
If the previous run is unfinished, skip this occurrence |
Deadline missed |
Continue |
Past the deadline, the transfer carries on |
The recommended value is Skip this run so the same files are not pulled twice. Skipping leaves the Flow's schedule intact. The next occurrence runs normally.
Schedule overlap offers two choices.
| Action | On an overlap | When to pick it |
|---|---|---|
Skip this run |
Drops this occurrence | The same files must not be pulled twice. The recommended value |
Queue one |
Queues this occurrence, which runs once the previous one finishes | No occurrence may be lost |
Queueing means a late run is followed by the one behind it, so on a job that slips every day the delay compounds. That is why the default is to skip.
Where the workspace changed this action, your Flow starts from that value instead. To see what yours is on, open Assurance and read the row: a
Workspacebadge means it came from the workspace defaults, aDefaultbadge means it is INNORIX's recommended value.
To change it, open the Assurance badge in the Flow editor → Error handling → Schedule overlap. See Decide in Advance What Happens When a Transfer Goes Wrong.
The schedule problems the product distinguishes#
Overlap is only one of them. Each situation is recorded under its own name with its own severity.
| Name | Code | When |
|---|---|---|
Schedule overlap |
SCH-01 | The previous run had not finished when the time came |
Manual and scheduled overlap |
SCH-03 | A manual run collided with the schedule |
Event burst |
SCH-04 | Events arrived in a rush and piled up run requests |
Missed schedule |
SCH-05 | The schedule never fired. Does not resolve itself |
Stale scheduled execution |
SCH-10 | A scheduled run is sitting in a stale state |
Late run |
SCH-11 | It started later than planned |
Scheduled start failed |
SCH-12 | The scheduled start itself failed. Does not resolve itself |
Run blocked by stuck run |
SCH-13 | A stuck run is blocking the next one |
Chain trigger lost |
SCH-14 | The signal linking stages was lost. Does not resolve itself |
Device offline at scheduled time |
OPS-03 | The device was off when the schedule fired |
Dependency not completed |
SCH-08 | A preceding job had not finished. A notice |
Policy changed during run |
SCH-09 | Policy changed mid-run. A notice |
Duplicate schedule merged |
SCH-02 | The same schedule arrived twice and was merged into one. A notice |
The three notices never appear in the issue list. They are left out of the run result and the issue counts, they do not show under Recovery, and they are not listed on their own. They are information, not failures.
So where do you see them? Their events are kept under Activity in the run detail. If it feels like something happened although there are no issues, check the event count under Activity.
The three that do not resolve themselves (Missed schedule, Scheduled start failed, Chain trigger lost) need a person. They appear under Unresolved in the Assurance report.
Handling a scheduled run by hand#
A scheduled run that has not started shows as Planned in Runs. Press the row, a detail panel opens on the right, and the ... in its header holds two things - the row itself has no ....
| Menu item | What it does |
|---|---|
Run now |
Runs immediately instead of waiting |
Skip this run |
Skips this occurrence only. The schedule stays |
A skipped occurrence becomes Skipped, and the same ... then offers only Run now.
Choosing by situation#
| Situation | What to do |
|---|---|
| Night collection sometimes slips into the morning | Keep Schedule overlap at Skip this run. Nothing is pulled twice |
| An aggregate that must never have a gap | Set Schedule overlap to Queue one. The late occurrence runs after the one ahead of it instead of being dropped |
| Maintenance today, skip just today | Skip this run from that day's Planned row |
| Urgent, needs to go before the schedule | Run now from the Planned row, accepting a possible Manual and scheduled overlap |
| Must avoid opening hours at an overseas site | Set the Time zone to the local one |
| Weekday-only business | Every weekday (Monday to Friday) |
Checking the result#
Calendar shows planned and past runs on a calendar, so the count per day is obvious.
In Runs, use the All statuses filter to look at Planned and Skipped separately.
In the Assurance report, check the schedule issues:
Issues & Recovery 1 recorded event
Activity 4 events · 1 policy decision1 policy decision on the Activity line means a policy fired once. A skipped overlap shows up here.
A clean outcome:
- The planned occurrence ended
Completed Issues & RecoveryreadsNo issuesContract StatusisSatisfied
Errors and exceptions#
| What you see | Meaning | What to do |
|---|---|---|
Missed schedule |
The schedule never fired | It does not resolve itself. Check the server and that the Flow is Active |
Scheduled start failed |
The start itself failed | Check device connections and paths, then run it manually |
Chain trigger lost |
The signal calling the next stage was lost | Check the previous stage, then run the next one manually |
Run blocked by stuck run |
A stuck run is in the way | Stopping that run releases the next |
Late run repeating |
It always starts late | Move the time earlier or split the targets |
Device offline at scheduled time |
The device was off | Review the device power policy, or set the action to Skip this device |
Event burst |
Events arrived in a rush | Narrow the event condition |
| Nothing runs at all | The Flow is Draft or Paused |
Turn it on with Save & activate |
Operating notes#
- A Flow must be
Activefor its schedule to run. Saving asDraftdoes not run. Check the status in the Flows list. - Shortening the interval without finding out why it slips makes overlap worse. When
Late runrepeats, split the targets or move the time. - Manual runs can collide with the schedule. Check the next scheduled time before running urgently.
- Do not schedule everything in head-office time. Overseas sites then transfer during their working day.
- Daylight saving is automatic. Do not adjust the time a second time.
- Put alerts on the three that do not resolve themselves. Delivered to Slack or Teams, nothing waits until morning.
By industry#
| Industry | How it is used |
|---|---|
| Manufacturing | Scheduled into the window between shifts, with time zone and time set to the plant |
| Retail | Out-of-hours rollouts. Every weekday (Monday to Friday) for weekdays, with weekend volume handled separately |
| Finance | Aggregation after close. No day may be missed, so overlap goes to Queue one |
| Healthcare | Scheduled around when inspection machines are off. Repeated Device offline at scheduled time means moving the time |
| Broadcast | Air time is the reference, so Late run is treated as an alarm |
| Public sector | The Policy decisions record documents why things were handled as they were |