SynapseHealer: Building an Autonomous, Self-Healing Interoperability DLQ with InterSystems IRIS AI Hub
Built for the InterSystems Programming Contest: AI Agents in Interoperability Workflows (2026) The problem: the DLQ is a graveyard In healthcare and supply-chain integrations, the dead-letter queue is where transactions go to die. A vendor renames client_uuid to customer_id with no warning. An EHR bridge emits 1985-02-31 , an impossible calendar day. An HL7 ADT^A01 arrives with PV1 before PID . A…
InterSystems IRIS AI Hub introduces SynapseHealer to address dead-letter queue (DLQ) issues in healthcare and supply-chain integrations. The DLQ serves as a repository for rejected transactions due to schema violations, impossible dates, or other anomalies. SynapseHealer replaces the manual intervention required to fix these issues with an automated pipeline.
The system consists of several components: Synapse.Process.DeadLetterHealerBP, Synapse.Agent.DeadLetterHealerAgent, Synapse.Agent.HealerToolSet, and a tri-branch router. The agent uses a governed %AI.Agent bound to Synapse.Agent.HealerToolSet, which has three specific calls: SearchRunbookRAG, ValidateRepairedPayload, and CalculateRiskScore. These calls help the agent look up verified runbooks, apply deterministic transformations, and validate the repaired payload before resubmission.
SynapseHealer defaults to the Gemini-1.5-flash model, which provides fast and cost-effective inference. The system tracks every inference in the Synapse.Data.AuditTelemetry table, storing prompt tokens, completion tokens, estimated cost, latency, tool-call trace, decision branch, risk level, and workflow status. The Next.js dashboard displays real-time cost and auto-heal rates.
Observability is achieved through a three-step trace, which includes SearchRunbookRAG, ValidateRepairedPayload, and CalculateRiskScore. These steps are displayed in the Next.js console, allowing humans to review the corrupted and healed payloads side-by-side. Approvals or rejections of healed messages are made through the workflow queue, with reviewer notes and queryable API endpoints available.
Finally, the IRIS agentic development process involves MCP build tooling, guardrails, ObjectScript syntax, and IPM packaging. Development-time tooling ensures the IDE and runtime governance work together seamlessly, while at runtime, the container governance is handled by native IRIS AI Hub functionality powered by an %AI.ToolSet. This setup enforces strict response contracts and validates repaired payloads to ensure data integrity.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
