Tell at a glance whether a Job is healthy

Your Jobs board is meant to answer one question in about two seconds: is anything here wrong, and if so, what do I do? Every Job card carries a small vocabulary that answers it — a status pill, a strip of coloured bars, a last-run line, and sometimes a button telling you the next move.

This page explains that vocabulary, so a red card sends you somewhere useful and a calm card genuinely means you can look away.

The strip of bars: how each day went

The seven bars on a Job card are its last seven days, oldest on the left. Open the Job and its Activity panel shows the same thing over fourteen days.

Colour tells you how much of the day went wrong — not merely whether anything did. A day where one run in four hundred failed shows a thin red line along the bottom of an otherwise green bar. A day where a quarter of the runs failed shows a red block roughly a quarter of the bar’s height. The two look different because they are different.

What you see	What happened that day
Solid green	Every run finished cleanly
Green with a thin red foot	A small share of runs failed
Mostly or entirely red	Most or all runs failed — a genuinely bad day
Amber band	Runs that finished but reported problems along the way, or were cancelled
Grey-blue	Runs still going when the page loaded
Pale grey	No runs, and none were expected
Amber wash	No runs, and that’s unusual for this Job — see Stopped running below

Bars are all the same height on the card, on purpose. A day with two runs and a day with two thousand get the same amount of space, so a quiet day is still big enough to see and tap. (On the Job’s own Activity panel, height does track volume — that panel has room to show both.)

Hover any bar for the exact numbers. You’ll get the date and a breakdown like “Aug 18 — 39 Succeeded, 1 Failed”. The colour is the summary; the tooltip is the truth. If you’re deciding whether to act, read the tooltip.

Why a single failure doesn’t turn the day red

It used to. The board coloured each day by the worst thing that happened in it, which sounds sensible until you run a Job a few hundred times a day: at that volume something fails on almost every calendar day, so the strip was red permanently and stopped meaning anything. Colour now tracks the share that failed, so a bad day actually stands out from a busy one.

The status pill: one problem, the most urgent one

Under the Job’s name sits a pill like Active · Failed last run. It shows the lifecycle status and, when something’s wrong, the single most urgent problem — not a list. Hover it for the specific reason.

Routario checks in this order and reports the first thing it finds:

Pill	What it means	Colour
Disconnected integration	The Job needs a connected service that isn’t connected any more	Red
Failed last run	The most recent run failed outright	Red
Stopped running	It’s active, but has gone quiet when it normally wouldn’t	Red
Overdue review	Someone was asked a question over 24 hours ago and hasn’t answered	Amber
Missing owner	Active, but nobody’s name is on it	Amber
Needs configuration	A draft that isn’t ready to activate yet	Amber
Paused	Deliberately stopped; it won’t run until you resume it	Amber
Healthy	Nothing to report	Green

The order matters. Because only the top match is shown, a Job that’s both unowned and broken tells you it’s broken — the more urgent fact wins. Assign an owner and any remaining problem surfaces next.

Red means something is broken and needs a person. Amber means something is incomplete or waiting. An unassigned owner is a gap to fill, not an emergency, so it’s amber — on a board you scan at a glance, everything shouting is the same as nothing shouting.

Stopped running: the quiet failure

The most expensive failure isn’t a red bar. It’s a Job that simply stops, with nothing to show for it.

Routario watches for that by comparing a Job against its own normal rhythm rather than a fixed rule. A Job that usually runs several times an hour is flagged after about two silent days. A Job that runs weekly isn’t flagged for going quiet over a weekend, because that’s just Tuesday for it.

When silence is genuinely unusual you’ll see three things at once: the days with no runs turn an amber wash instead of empty grey, the card says “No runs in 5 days · typically 40/day” in plain words, and the pill reads Stopped running. Usually the cause is a trigger — a schedule switched off, a webhook that stopped arriving, an inbox that stopped matching.

It needs history first

A Job has to be around eight weeks old before Routario has enough of a pattern to call silence unusual. Below that it stays quiet rather than guessing, so a freshly installed Job won’t cry wolf. Paused Jobs are never flagged either — they’re quiet because you made them quiet.

The rest of the card

The last-run line — a coloured dot, a relative time, and the outcome in words: “3m ago · Succeeded”. The words are there deliberately, so the meaning never rests on colour alone.

The next-step button, bottom right, when there’s an obvious move:

Button	Appears when	Takes you to
View failure	The last run failed	The run, so you can read the error
Check trigger	The Job has stopped running	Its configuration, to see what stopped firing it
Review	Someone’s waiting on an answer	Your Inbox, filtered to this Job
Assign owner	Active with nobody’s name on it	The owner picker, already open
Set up / Activate / Resume	It’s a draft, or paused	The matching next step

No button means there’s nothing obvious to do — which, on a healthy Job, is the correct answer.

A worked example

You open the board on a Monday and see three cards.

“Invoice intake” — green bars all week with a hairline of red along two of them, pill reads Active, last run “12m ago · Succeeded”. Hovering Thursday’s bar says “41 Succeeded, 1 Failed”. One invoice in forty-two failed and the rest went through. Nothing is on fire; if you care about that one invoice it’s in the Logs, but the Job is doing its job. Leave it.

“Ticket triage” — the first two bars are green, the last five are an amber wash, and the card says “No runs in 5 days · typically 30/day” with a Stopped running pill and a Check trigger button. This is the one to open. Something stopped feeding it on Wednesday and roughly 150 tickets a day have gone untriaged since. This is your morning.

“Weekly digest” — mostly pale grey with one green bar, pill reads Active, last run “1d ago · Succeeded”. Grey here is correct: it runs once a week and it ran. Leave it.

Two seconds, three cards, one of them your actual problem. That’s what the vocabulary is for.

Where to go next

• Keeping your automations healthy — the agent that reads these same run outcomes on a schedule and proposes fixes before you notice them
• See what happened — the Logs page — the run-by-run detail behind every bar on the strip
• Jobs — what a Job is and how its parts fit together