Logic Apps Subflows not loading for a period of time

Hi @Christopher Fryett

Thank you for reaching out to Microsoft Q&A.

The issue you are experiencing occurs because the Logic Apps Standard designer must fully load workflow metadata, referenced workflows, API connections, and parameter schemas before the designer becomes stable. When you have 60+ workflows, the designer performs deep introspection on every referenced child workflow. This leads to slow parameter hydration, the spinner you see beside actions that call other workflows, and situations where the designer allows a save before all components have fully loaded. This results in missing or uninitialized parameters, especially in workflows using database connectors. Additionally, because your Logic Apps Standard instance runs in an ASE, file system and metadata retrieval operations incur additional latency, further degrading performance. Although I searched your organization’s files, meetings, emails, chats, and events, there is no internal documentation available that addresses Logic Apps Standard designer performance or parameter loading delays.

Refer below points to resolve this issue or use the following workarounds:

1. Use Visual Studio Code instead of the Portal Designer

The portal designer performs heavy metadata hydration, which slows down as your workflow count grows. VS Code works directly with the underlying JSON files and avoids this delay. This approach also aligns with internal Microsoft-directed development practices referenced in your organization’s notifications about Logic Apps development using VS Code.

2. Break your Logic Apps solution into smaller Standard projects

Avoid placing dozens of workflows into a single Logic App Standard project. Split workflows into grouped projects (e.g., related functional domains). This significantly reduces designer load time because metadata hydration becomes lighter per project.

3. Avoid relying on “child workflow” designer actions for large-scale solutions

Whenever possible, consider switching child-workflow calls to HTTP actions or Service Bus–based orchestration, which reduces dependency on the designer loading deeply nested workflow parameter definitions.

4. Reduce designer UI rendering load

Collapse scopes and avoid opening multiple browser tabs containing designer sessions to reduce initialization time and prevent partially loaded UI states.

5. Use stateless workflows wherever applicable

Stateless workflows initialize faster and contain less metadata to parse compared to stateful workflows. Although the training deck found in your organization does not explicitly mention designer performance, it does document the difference between stateful and stateless workflows, validating that stateless workflows are lighter.

This behaviour is consistent with Logic Apps Standard designer metadata loading overhead rather than a workflow execution issue.

In Logic Apps Standard running inside an ASE plan, the designer loads workflow definitions, connector metadata, parameter schemas, and dependent workflow references at design time. When a workflow calls other workflows using the built in Workflow action, the designer must resolve:

• Referenced workflow definition

• Input and output schemas

• Parameter metadata

• Connector configuration bindings

With more than sixty workflows in the same Logic App project, the metadata graph becomes large. Each time the designer opens a workflow, it enumerates all local workflows and resolves references. Database connectors increase the impact because their OpenAPI schemas and dynamic content definitions are also retrieved and parsed.

This explains the spinning indicator on the action card. The designer has not fully initialised the action metadata. If you save before initialisation completes, parameter values may not be fully bound in the JSON definition, which leads to incomplete saves.

Key contributing factors in Logic Apps Standard:

1. Large number of workflows in a single Logic App resource
2. Heavy connector usage with dynamic schemas
3. ASE networking latency
4. App Service plan resource pressure
5. File system latency in the underlying content share

In Standard, all workflows are stored as files in the same project structure. The designer reads these files and builds an in memory model. As the project grows, this operation slows down.

Recommended mitigation approach:

First, validate that this is designer overhead and not runtime degradation. Check:

• Workflow execution duration in Monitor

• App Service CPU and memory metrics

• Storage latency for the content share

If runtime performance is healthy but designer is slow, focus on architectural segmentation.

Best practice for Logic Apps Standard at scale is to distribute workflows across multiple Logic App resources rather than keeping dozens inside a single project. For example:

• Group workflows by domain or bounded context

• Separate database heavy workflows

• Isolate reusable child workflows into a dedicated Logic App

This reduces metadata resolution time in the designer.

You should also review your App Service plan sizing. In ASE, ensure the worker instance has sufficient CPU and memory. Designer rendering depends on the same worker resources as runtime.

Additional optimisation steps:

• Avoid unnecessary dynamic content binding

• Replace some dynamic schemas with fixed JSON schemas where possible

• Reduce nested workflow depth

• Periodically clean unused workflows from the project

If you are using an Azure Files content share, verify latency and IOPS. High file system latency directly impacts Standard designer performance because definitions are read from disk.

This is a known scaling characteristic of Logic Apps Standard rather than a configuration error. The most effective long term solution is logical separation into multiple Logic App resources to keep each project lightweight and maintain predictable designer responsiveness.

If you can confirm the number of connectors used and the SKU of the App Service plan inside the ASE, further tuning guidance can be provided.