Integration Watchdog

When two of your systems stop talking nothing beeps, so now something does, the hour it happens.

Orders stop importing. Payments stop reaching the books. The customer list stops updating. Everything looks completely normal until somebody says the numbers look wrong — and by then it has usually been days, and the missing records have to be found and re-entered by hand.

Your Integration Watchdog watches the connections themselves. Every hour it asks each one a single question: has anything actually come through since it normally would have?

It judges each connection by its own habit. A connection that moves records every fifteen minutes and has moved none in four hours is broken. A nightly one that has moved none in four hours is fine. A shared threshold cannot tell those apart, which is why generic monitoring either cries wolf every week or misses the real thing.

It knows quiet from broken. Before calling anything a failure it asks whether anything should have come through at all — a holiday, a closed shop, or the other company's own outage looks identical in a log. False alarms every week are how the real alarm gets ignored.

It catches the credential before it lapses. The most common cause of a silent outage is a token that expired over a weekend. Once a week it lists every token, secret and certificate with a date on it, says who can renew each one, and says what stops working when it goes — early enough that renewing it is routine rather than an emergency.

It tells you what stopped, not what errored. Not "sync job 4 has not run since Thursday" but "payments have not reached the ledger since Thursday, so this week's figures are short".

It is careful about retrying. One retry on a transient failure. Never a retry storm against a system that writes real records, and never an automatic retry at all on anything carrying money.

Connects to your integrations. Works alongside your help desk and whoever owns each connected system.