Workflow break check · your result
Where your workflow between your tools typically breaks
These are the places setups like yours usually fail; they are not findings about your systems, since nothing here looked at them. Each one comes with a way to check it yourself. Navceed is being built to run checks of these six kinds; today it runs them on an example company, and connecting your own tools is coming next.
1. The counts at each end quietly differ
Check firstAnywhereYou can run thisabout 45 min
Shown because you answered “Never”.
How to check, and why it happens
The number that left one tool and the number that arrived in the next drift apart, and no report shows the gap because each tool only reports its own side.
Why it happens. Counts are read when they are reported, not compared as they happen.
Usually fixed by: Whoever owns the reporting.
Kind of break, one of six: Count: do the numbers at each end agree?
No public source found for this one: it is our reading of how these hand-offs usually work.
2. A step stops, or runs late, and nobody is told
Check firstAnywhereRun by whoever runs the stepabout 20 min
Shown because you answered “Someone would tell us”.
How to check, and why it happens
A step that should run every hour or every day stops, or slips by hours, while everything around it looks normal. Work piles up behind it.
Why it happens. A connection lost its login, hit a limit or failed once and was never retried, and an error log nobody reads is the only place it shows.
Usually fixed by: Whoever runs the step.
Kind of break, one of six: Freshness: is the step still running, and on time?
No public source found for this one: it is our reading of how these hand-offs usually work.
3. A rule that should always hold is not checked
Check firstAnywhereRun by the tool's adminabout 30 min
Shown because you answered “It is a rule on paper only”.
How to check, and why it happens
There is a rule everyone believes in (every record has an owner, no duplicates, nothing waits more than a day) and nothing checks it, so exceptions pile up unseen.
Why it happens. The rule was agreed once and nobody turned it into a check that runs on its own.
Usually fixed by: The owner of the rule.
Kind of break, one of six: Condition: does a rule that should always hold, hold?
No public source found for this one: it is our reading of how these hand-offs usually work.
Show more (1)
4. A record never arrives in the next tool
Worth a quick lookAnywhereYou can run thisabout 20 min
Shown because you answered “I'm not sure, to: How would you find out if a record never arrived in the next tool?”.
How to check, and why it happens
Something that should move from one tool to the next is refused, skipped or held back, and the person or company it was about never shows up. The first tool says it sent it.
Why it happens. A rule, a required field, a daily limit or a lost login stops it, and nothing tells the people who care.
Usually fixed by: Whoever owns the connection, with the owner of the receiving tool.
Kind of break, one of six: Arrival: did it show up in the next tool in time?
No public source found for this one: it is our reading of how these hand-offs usually work.
The common questions follow six kinds of check: Arrival (did it show up in the next tool in time?), Return path (does what the next tool did get back to the first?), Count (do the numbers at each end agree?), Freshness (is the step still running, and on time?), Field landed (did each field arrive with a value, in the right place?), Condition (does a rule that should always hold, hold?).
Checking by hand does not last
Everything above is a one-off look. The breaks that cost you are the ones that start next Tuesday, when someone changes a field.
No sign-up was needed for this page and nothing you chose was saved. Print this page to keep it. Start again.