Azure API Management
An Azure service that provides a hybrid, multi-cloud management platform for APIs.
2,452 questions
MCP Server API Fails on Higher Tiers in Azure APIM: "Client connection was unexpectedly closed"
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(156.8, 42%, 25%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3EMS%3C/text%3E%3C/svg%3E)
M., Sai Kiran
0
Reputation points
I'm exposing a REST API as an MCP server using Azure API Management. It works fine on the Basic tier — the tools are fetched correctly, and both GET and POST MCP calls succeed. However, when I use the same setup on Standard or Premium tiers, it fails. The client can't fetch the tools, and APIM logs show the below error:
"reason":"ClientConnectionFailure","message":"Client connection was unexpectedly closed.","section":"outbound".
The backend API is accessible directly and has no firewall or certificate issues. Why does this happen only on higher tiers, and how can I fix or debug this behavior?
Azure API Management
{count} votes
-
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(281.6, 21%, 37%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3EKC%3C/text%3E%3C/svg%3E)
Krishna Chowdary Paricharla 280 Reputation points Microsoft External Staff Moderator2025-07-03T07:53:15.9433333+00:00 Hello M., Sai Kiran •
Thanks for reaching out!
The "Client connection was unexpectedly closed" error on Standard or Premium APIM tiers often results from networking differences compared to the Basic tier. Consider these points:
- VNet Integration: Standard and Premium tiers may route traffic through a VNet. Check that outbound connectivity, DNS, NSGs, and firewalls are set up properly.
- TCP Idle Timeouts: Long-running requests might be interrupted due to idle timeouts. Use keep-alive options or adjust backend settings as needed.
- TLS/SSL Issues: Higher tiers can require stricter TLS settings. Make sure your MCP server supports the necessary TLS versions and ciphers.
- Header or payload size limits may be higher in these tiers, so verify your requests fit within the limits.
Recommended steps:
- Turn on diagnostic logs and network tracing in APIM.
- Test backend connectivity from within the same subnet.
- Compare with an external APIM instance to help identify the problem.
For your reference, please go through this following documentation:
Hope this helps!
Sign in to comment
Sign in to answer