OIC comes with a monitoring dashboard that shows integration instance status, error counts, and message payloads. It works. But for production environments with multiple integrations running continuously, you need more than a dashboard someone has to actively check.
What the default monitoring gives you
- Instance-level status (succeeded / failed / aborted)
- Error messages and fault details
- Message payload inspection (if tracking is enabled)
- Basic retry capability
This is useful for debugging individual failures. It’s not a monitoring strategy.
What production monitoring actually needs
Proactive alerting — you should know about a failed integration before a business user does. OIC supports email notifications on failure — configure these for every business-critical integration. Route alerts to a shared mailbox, not an individual.
Failure rate trending — a single failure in a scheduled integration might be noise. Five consecutive failures is a pattern. Build a simple tracking mechanism (an ATP table updated by a monitoring integration) that can detect degrading reliability before it becomes an outage.
Heartbeat checks — for integrations that run on a schedule, a missed run is as bad as a failed run — and harder to detect. After each scheduled run, post a heartbeat signal to a lightweight endpoint. If the heartbeat stops arriving, trigger an alert.
OIC notification configuration
Navigate to: Settings → Notifications in the OIC console. Configure:
- Email recipients for failed instances
- Threshold for notification (immediate vs. after N failures)
- Include payload information where it doesn’t contain sensitive data
Building a monitoring integration
A simple monitoring pattern:
Monitoring Integration (scheduled every 15 mins):
1. Query OIC Management REST API for failed instances in last 15 mins
2. If count > threshold: send email notification with instance IDs
3. Log summary to tracking table
OIC exposes a Management REST API that lets you query instance status programmatically. This is how you build cross-integration visibility without manual dashboard checking.
Payload tracking
Enable message tracking for business-critical integrations. Store the tracking fields (e.g. source transaction ID, document number) — not the full payload — in tracked fields. This makes it possible to search for a specific transaction by ID when a business user reports an issue.