# Idle Azure-SSIS Integration Runtime Left Running Canonical: https://www.pointfive.co/efficiency-hub/inefficiencies/idle-azure-ssis-integration-runtime-left-running An Azure-SSIS integration runtime (IR) is a dedicated cluster of VM nodes that Azure Data Factory or Synapse pipelines use to run lifted-and-shifted... By: PointFive Updated: 2026-09-28 [Cloud Efficiency Hub](https://www.pointfive.co/efficiency-hub) The short version An Azure-SSIS integration runtime (IR) is a dedicated cluster of VM nodes that Azure Data Factory or Synapse pipelines use to run lifted-and-shifted SQL Server Integration Services packages. PointFive Research Cloud cost research at PointFive Azure service [Azure Data Factory V2](https://www.pointfive.co/efficiency-hub/cloud-services/azure-data-factory-v2) Category [Compute](https://www.pointfive.co/efficiency-hub/service-category/compute) Reference CER-0450 Type Idle or Unused Resource ## Explanation Why the waste happens and who it affects. Once started, it bills for every node for as long as it stays started, whether or not any package is executing. Many SSIS workloads run in short nightly or hourly windows, yet the IR is started once during migration and left running around the clock. Unlike the rest of Data Factory, which is billed per activity run and per unit of data movement or data flow compute, the SSIS IR behaves like always-on infrastructure. Microsoft's guidance is explicit: running an Azure-SSIS IR has a cost, so it should run only when SSIS packages need to run, and Data Factory provides a pipeline template that starts the IR just before a package and stops it right after so it doesn't run idly. The waste is easy to miss because, even with per-pipeline billing enabled, SSIS node charges file under a factory-level fall-back line item rather than under the pipelines that use them. ## Billing model The pricing dimensions that drive this cost. SSIS IR node Billed per node in per-second increments while the IR is started, based on VM size and SSIS edition Standard and Enterprise editions Priced separately per node; Enterprise edition is needed only for Enterprise-only SSIS features Azure Hybrid Benefit Reduced node pricing for customers with SQL Server licenses covered by active Software Assurance Stopped IR Stopping the IR releases all of its nodes and stops billing; the IR must also be stopped before it can be reconfigured or deleted ## How to detect 4 checks to find it in your estate. - List Azure-SSIS integration runtimes and their state (Started or Stopped) across factories and Synapse workspaces, for example with Get-AzDataFactoryV2IntegrationRuntime or the integration runtimes REST API - Compare the hours each IR spends in the Started state with the time Execute SSIS Package activities actually run, using pipeline and activity run history in the Monitor tab - Check Cost analysis for the SSIS node meters under Azure Data Factory v2; charges that accrue 24x7 for an IR whose packages run only in short windows indicate idle runtime - Review node count, node size and edition against package concurrency and SSIS features used, since oversized or Enterprise nodes multiply idle cost ## How to fix 5 ways to remove the waste. - Chain Execute SSIS Package activities between web activities that start and stop the IR, or use the 'Schedule ADF pipeline to start and stop Azure-SSIS IR just in time before and after running SSIS package' template - For fixed ETL windows, schedule separate start and stop pipelines with triggers, or use an Azure Automation runbook with Start-AzDataFactoryV2IntegrationRuntime and Stop-AzDataFactoryV2IntegrationRuntime - Use Until activities that check the IR state so start and stop steps retry on transient errors and complete only when the IR is actually Started or Stopped, and allow for the IR's startup time in the schedule - Rightsize node size and count to the packages' concurrency, compare Standard and Enterprise edition node prices and use Standard unless Enterprise features are required, and apply Azure Hybrid Benefit where eligible SQL Server licenses exist - Delete SSIS IRs left over from completed migrations ## Documentation Vendor references for pricing and configuration. - [Schedule an Azure-SSIS integration runtime learn.microsoft.com](https://learn.microsoft.com/en-us/azure/data-factory/how-to-schedule-azure-ssis-integration-runtime) - [Plan to manage costs for Azure Data Factory learn.microsoft.com](https://learn.microsoft.com/en-us/azure/data-factory/plan-manage-costs) - [SQL Server Integration Services (SSIS) Pricing - Data Factory azure.microsoft.com](https://azure.microsoft.com/en-us/pricing/details/data-factory/ssis/) - [Reconfigure the Azure-SSIS integration runtime learn.microsoft.com](https://learn.microsoft.com/en-us/azure/data-factory/manage-azure-ssis-integration-runtime) ## Related inefficiencies [Browse the library](https://www.pointfive.co/efficiency-hub) - Azure Data Factory V2 CER-0261 ### [Suboptimal Integration Runtime Region Selection in Azure Data Factory](https://www.pointfive.co/efficiency-hub/inefficiencies/suboptimal-integration-runtime-region-selection-in-azure-data-factory-0bcad) When Integration Runtimes use the default "Auto Resolve" region setting, the work does not always run in the region where the data lives. For copy activities, Azure makes a best effort to run in the sink data store's region (or the closest... Networking - Azure Functions CER-0315 ### [Orphaned Azure Function Apps with No Active Functions or Triggers](https://www.pointfive.co/efficiency-hub/inefficiencies/orphaned-azure-function-apps-with-no-active-functions-or-triggers) Azure Function apps can persist long after the applications or workflows they supported have been retired - particularly in development, testing, and experimentation environments where cleanup is often overlooked. Even when no functions... Compute - Azure AKS CER-0099 ### [Orphaned Kubernetes Resources in AKS](https://www.pointfive.co/efficiency-hub/inefficiencies/orphaned-kubernetes-resources-230dd) Kubernetes environments often accumulate unused resources over time as applications evolve. Common examples include Persistent Volume Claims (PVCs) backed by Azure Disks, Services that trigger load balancer provisioning, or stale... Compute --- Source: the public page above. Product screenshots and illustrative interfaces are examples, not live customer data.