OIC Error Hospital: Safe Resubmission Strategies

The OIC Error Hospital

The Error Hospital is OIC’s centralized dashboard for failed integration instances. Every instance that faults ends up here, grouped by the error type and integration.

flowchart TD
    INST[Integration Instance] --> RUN{Execution}
    RUN -->|Success| DONE[✓ Completed]
    RUN -->|Fault| ERR[Error Hospital]
    ERR --> INV[Investigate root cause]
    INV --> DEC{Decision}
    DEC -->|Fix required first| FIX[Fix data/config/system]
    FIX --> RESUB
    DEC -->|Transient error| RESUB[Resubmit instance]
    DEC -->|Discard| ABORT[Abort instance]
    RESUB --> RUN

    style DONE fill:#0f2d1f,stroke:#22C55E,color:#e2e8f0
    style ERR fill:#2d0f0f,stroke:#EF4444,color:#e2e8f0
    style ABORT fill:#1e293b,stroke:#475569,color:#94A3B8

Safe resubmission process

Before resubmitting, always verify:

checklist

1. Identify the exact failure point from the instance activity trace
2. Check downstream system state — did any partial writes occur before the fault?
3. If partial writes occurred — manually reconcile or compensate first
4. Resubmit a limited batch — test with 1 instance before bulk resubmitting

Configuring automatic retry

The safest approach: configure retry logic in the integration itself so the error hospital is rarely needed:

Integration trigger configuration:
  On Fault:
    Retry Count: 3
    Retry Interval: 30 seconds (exponential: 30s, 60s, 120s)
    Retry On: HTTP 5xx, Connection Timeout
    No Retry On: HTTP 4xx (bad data — retry won't help)

For the exponential backoff in a custom error handler:

// JavaScript action in fault handler
var attempt = $input.retryCount || 0;
var baseDelay = 30;
var delay = baseDelay * Math.pow(2, attempt); // 30, 60, 120, 240...
return {
  shouldRetry: attempt < 3,
  delaySeconds: Math.min(delay, 300), // cap at 5 minutes
  nextAttempt: attempt + 1
};

Bulk resubmit via OIC REST API

For resubmitting many failed instances at once — don’t click through the UI:

# Get all failed instances for an integration
curl -u "$OIC_USER:$OIC_PASS" \
  "https://$OIC_HOST/ic/api/integration/v1/monitoring/instances?\
    integrationId=INVOICE_SYNC|01.00.0000&status=FAILED&limit=100" \
  | jq '.items[].id' > failed_instance_ids.txt

# Resubmit each
while read id; do
  curl -u "$OIC_USER:$OIC_PASS" \
    -X POST \
    "https://$OIC_HOST/ic/api/integration/v1/monitoring/instances/$id/resubmit"
  echo "Resubmitted: $id"
done < failed_instance_ids.txt

Automation opportunity

For integrations with predictable transient failure patterns, configure OIC’s retry capability — exponential backoff retries before the instance is marked as failed. This handles the majority of transient errors without any manual intervention or Error Hospital visits.