Why Zero-Downtime Deployments Matter for Integration
Integration workflows are often business-critical — an order management flow, a payment notification, a stock update. Deploying a new version by swapping out the running app introduces a window of downtime that directly impacts business operations. For enterprise clients, even a 30-second gap can mean dropped messages and manual recovery effort.
Logic App Standard (the single-tenant, App Service Plan–based tier) supports deployment slots — the same blue/green deployment model used by Azure App Service. This guide explains how to implement it correctly.
Deployment Slot Architecture
The pattern is straightforward:
- Production slot — the live, traffic-serving deployment
- Staging slot — receives your new deployment, runs smoke tests
- Swap operation — atomic swap of staging → production with instant rollback capability
During the swap, Azure warms up the staging slot before cutting over traffic. Connections are drained from production gracefully.
The Managed Identity Challenge
The tricky part of slot swaps for Logic Apps is connections. Logic App Standard stores connection references in connections.json — and each slot has its own Managed Identity with its own RBAC assignments.
The solution is to use parameterised connections that reference Key Vault secrets or App Settings — and ensure both slots have their Managed Identity granted the necessary roles on the same downstream resources.
Azure DevOps Pipeline Structure
A production-grade pipeline for Logic App Standard deployment:
- Stage 1: Build — package workflow definitions, parameters files, connections.json
- Stage 2: Deploy to staging —
az logicapp deployment source config-ziptargeting the staging slot - Stage 3: Smoke test — trigger test workflows, validate outputs via Azure Monitor
- Stage 4: Slot swap —
az webapp deployment slot swappromoting staging to production - Stage 5: Post-swap validation — verify production is healthy; auto-rollback on failure
Variable Group Promotion
Use Azure DevOps Variable Groups scoped per environment (Dev, Test, Prod). Each group holds environment-specific values — connection strings, endpoint URLs, feature flags. The pipeline promotes the same artifact through environments, only the variable group changes at each stage.
This is the deployment pattern we use for all Neuaxis Logic App Standard deliveries, including the BIOVIA OneLab project. Read the case study →
Need help implementing this?
Neuaxis delivers enterprise integration solutions — Azure iPaaS, BizTalk, MuleSoft, Kafka and Neuron ESB.
Book a Free Consultation ↗