MonitoringFind out before your client does

Feeds break quietly. A supplier file arrives half-empty, a price column stops filling, an overnight import fails — and nobody notices until the ads are already wrong. A monitor is a condition you write about your own numbers, watching every catalog and channel you run. When it trips it comes and finds you — and it can stop the bad feed from reaching your ads at all.

You write the conditionOne monitor, every accountEmails the right peopleCan hold the publish
Start freeUmów demo
Monitoring
Triggered on Maison Vertewatching 12 catalogs

After each catalog import on all 12 catalogs — when product count drops more than 25% — raise a Warning alert, email 2 people, and hold the publish.

Maison Verte lost 41% of its products
Import 12 minutes ago · publish held · the previous feed is still live
Open catalog

You decide what counts as wrong

Anyone can tell you the product count dropped. That’s the break you were going to notice anyway. The ones that quietly cost you money are narrower — the field that stopped filling, the supplier column that went stale, the third of the catalog that went out of stock overnight. So the condition is yours to write: pick the check, pick the statistic, pick the operator, set the number. Down to a single named field.

Monitor · Key fields stopped filling
On
Check after each catalog import

When

Fill rate % of sale_price is less than 60orFill rate % change of gtin is less than −5

Then

Raise a Warning alertEmail the ops inbox

No vendor picked that condition for you, and no threshold was guessed on your behalf. Combine a few with and / or when one number alone would be too jumpy.

Statistics you can build on

After each catalog import

Product count, and how it changed
New, removed and excluded products
Products that couldn’t be read, and products with warnings
Fill rate % of any field — and how that rate changed
% of products whose any field changed

After each channel feed update

Product count change %, and excluded products
Out-of-stock products, and out-of-stock %
Whether the publish failed

On a schedule

Hours since the last successful import

One monitor. Every client you run.

Set the rule once and point it at everything — every catalog, every channel, or just the handful you’re nervous about. Each one is judged on its own, so a monitor watching forty accounts can be shouting about one and quiet about the other thirty-nine.

Monitor · Product count drops sharply
On
Check after each catalog import

When

product count change is below −25%

Then

Raise a Warning alertEmail the client, and the ops inboxHold the publish
Watching
All catalogs · 12
Northwind Homeall clear
Alpine Outfittersall clear
Maison Vertetriggered
Coastline Kidsall clear

+ 8 more

Add a client next month and the monitor picks it up. No rule to copy twelve times.

Four moments to check

A monitor runs after each catalog import, after each channel feed update, on a regular schedule for the conditions that are about time, or whenever an event is reported. You pick the moment; the statistics on offer change to match it.

Watches catalogs and channels

A monitor on a catalog catches the problem where it enters — and holds every channel that exports those products, before the bad data reaches any destination. A monitor on a channel watches what happens on the way out, and holds just that one.

Already watching on day one

Every account starts with three monitors switched on: a failed import, a failed publish, and a catalog that loses more than 30% of its products between imports. Adjust the thresholds, add email, or leave them be — they work before you configure anything.

An alert that waits to be found isn’t an alert

A status that turns red on a page helps the person who happens to open that page. When you have thirty clients and half an hour each, that person is nobody. A monitor comes to you instead — and one monitor can send several emails, each with its own wording and its own recipients: team members, or outside addresses that confirm themselves first. The client wants to hear that you’re on it. Your ops inbox wants the numbers.

Email alert
alex@northwindhome.com

A heads-up about your product feed

Today’s update from your store had far fewer products than usual, so we’ve paused it — your ads are still running on yesterday’s products. We’re looking into it and will confirm once it’s back to normal.

Email alert
feeds@youragency.com

Northwind Home · product count −41%

Catalog: Northwind Home. Import at 06:12 returned 2,420 products, down from 4,102. Meta publish held; previous feed still live.

In the app, where you already look

Every alert lands on your dashboard’s attention list and on the catalog or channel it came from — the same place your warnings and errors already live. Email is the extra, not the only copy.

One email per incident

A monitor emails when it triggers — not on every check while the problem lasts. Nobody learns to filter your alerts into a folder.

Alerts clear themselves

An alert isn’t a ticket to close. When the next run looks healthy it disappears on its own, so the ones still sitting there are exactly the ones that still matter.
See what Renable watches inside the feed

Describe what would ruin your week. She builds the monitor.

You know what you’re afraid of long before you know which statistic expresses it. Say it in your own words and Aimée turns it into a monitor — which check, which threshold, who gets told. Every one of them opens up afterwards, so you can read it, tune it, or build the next one by hand.

Aimée

Tell me if any client’s product count falls off a cliff — and don’t let a feed like that go out to Meta.

Created monitor: Product count drops sharply

After each catalog import on all 12 catalogs — when product count drops more than 25% — raise a Warning alert and hold the publish.

Drafted it — nothing’s watching yet. Who should the alert email go to, and should I publish the monitor?

PublikujLet me look first

Nothing happens until you publish

A monitor Aimée writes is a draft, like every other change she makes. You read the sentence it turns into, then publish it — and pause it with one switch if it ever gets noisy.

Or build it yourself

The whole thing is a normal page: pick the check, set the condition, choose what happens. Aimée is the faster route, never the only one.
Meet Aimée

And when it matters, it can stop the feed

Knowing your feed broke is worth something. Not shipping it is worth a lot more. Give a monitor the power to hold a channel’s publish, and a suspicious feed simply doesn’t go out — the last good feed keeps serving your ads while you look into it.

Channel · Northwind Home · Meta
Publish held

Feed updates

Today 06:12 · Publish held2,420 products
Yesterday 06:12 · Published — live now4,102 products

Meta keeps reading yesterday’s feed. Nothing went dark, and nothing wrong went out.

Publish · Northwind Home · Meta

Product count drops sharply is holding this publish.

The previous feed stays live until a check passes. If you’ve looked at the data and it’s genuinely fine, you can send this update anyway.

Publish anywayCancel

Your ads never go dark

A held publish doesn’t empty anything. The feed the ad network already has keeps serving, so campaigns keep running on the last data you trusted.

It clears itself

The next import or feed update that looks healthy publishes normally and the alert disappears. Fix the cause and the hold lifts on its own — there’s no queue to release and nothing to remember to undo.

You always have the override

Sometimes the catalog really did halve because a season ended. Publish anyway, and that update goes live — the monitor keeps watching, so it still catches the next one.

Stop hearing about broken feeds from your clients

Connect a feed and the first monitors are already watching it. Monitoring, alert emails and the publish hold are on every plan, including the free one — no add-on, no upgrade to be told your feed broke.

Start freeUmów demo