Production Monitoring: Your Processes Monitored Continuously

Infrastructure monitoring tracks components such as server availability, URL response times, and error rates. Workflow monitoring simulates the entire business scenario. This is what makes it possible to detect a functional failure even when the entire technical stack is showing "green."

You choose the frequency on a per-scenario basis, ranging from a cycle of a few minutes for a payment flow to a daily run for a secondary form. The appropriate setting is determined by the cost of one hour of downtime for the flow in question.

The alert is sent via the channel defined for the scenario: email, Slack, SMS, or a webhook to your own tool. Critical scenarios can have a different channel, recipient, and time limit than other scenarios, which prevents the on-call team from being woken up for a minor incident.

A failure is first automatically replayed: anything caused by a temporary glitch disappears without triggering an alert. Anything that persists after replay triggers an incident, and failures with the same root cause are grouped into a single incident rather than generating an alert for each occurrence. A human then assesses the incident before it becomes an alert.

The rule is simple: in production, we use only data created for testing. The test scenarios rely on dedicated accounts and datasets, and any steps that would produce an undesirable real-world effect are limited to the scope of the test scenario.

Yes, and it’s the fastest way to get started. A pre-written non-regression test scenario becomes a production probe by changing its environment and data, and the same scenario is also used to measure response times step by step. You only need to maintain a single test suite, whereas a testing tool, a load testing tool, and a monitoring tool would require you to maintain three.

No, and it’s best to establish this limitation right from the start. Mr Suricate synthetic monitoring: it replays complete user journeys at regular intervals. It does not measure your visitors’ actual sessions and does not instrument your application code.

End-to-end user journeys, not just page availability. A website might return a 200 status code but have a broken checkout process at the payment stage—this is exactly the kind of issue that journey monitoring detects.

Executions can be identified and excluded from your reports by IP address or by parameter, depending on your tool. The filtering is set up at launch, before the first campaign.

Dedicated test accounts are sufficient in most cases. For sensitive steps, we work together to define the scope of the scenario and what it does not trigger.

Anything that appears in an interface: a lead time, a price, inventory levels, a time slot, a fee, or a required disclosure. If a customer can see it, we can note it and compare it to what is expected.

No, and we prefer to be upfront about it. We provide a detailed business impact analysis: which user journey is affected, at what stage, and since when. The cost estimate in euros depends on your conversion data and is calculated in collaboration with your analytics team.