Business Central telemetry and monitoring

Business Central emits detailed telemetry — every page load, API call, posting routine, error, and platform event — to Microsoft's Azure Application Insights. Partners and customers wire their tenants into their own Application Insights workspaces and use the data to monitor health, diagnose incidents, and improve their extensions.

Setup

Configure an Application Insights instrumentation key on the tenant (from the BC admin centre) for tenant-wide telemetry, on individual environments for environment-scoped telemetry, and inside the app.json of each extension for extension-scoped streams. The three layers can target the same workspace or different ones.

What's emitted

Microsoft documents dozens of standard telemetry events: page loads, report runs, job queue executions, API calls (with response time and status), authentication events, permission errors, AL runtime errors, lock timeouts, slow queries, database storage thresholds, environment lifecycle events (update started, completed, failed), and extension lifecycle events. Each event carries structured properties: tenant ID, environment, user, duration, exception details where applicable.

Querying

Application Insights uses Kusto Query Language (KQL). A typical analysis query: filter customEvents to a date range, group by environment and event name, project p95 duration. Microsoft publishes a library of canned queries (the BCTech GitHub repo) for the most common questions: which pages are slow, which integrations time out, which users hit permission errors.

Workbooks

Application Insights workbooks turn KQL queries into shareable dashboards. Microsoft publishes a Business Central workbook with pre-built views of health, performance, and errors. Most partners customise it.

Alerts

Application Insights alerts fire on thresholds — error rate over 5%, lock timeouts more than N per minute, an environment failing to update — and route to Teams, email, or Azure Monitor action groups. Set them up before you need them.

Custom telemetry

Your AL extensions can emit their own telemetry via Session.LogMessage with a verbosity level (Verbose, Normal, Warning, Error, Critical). This is how you trace your own code's behaviour in production.

Retention and cost

Application Insights retains data for a configurable period (default 90 days), and charges by ingested volume. For high-volume tenants, sample verbose events and keep the structural ones.

Operational reality

A partner with a fleet of customer tenants and good telemetry can spot a release wave breaking change before the customer feels it. A partner without telemetry finds out by phone call. The investment is small; the leverage is large.