Azure Logic Apps jobs fail prematurely despite Failure Interval configuration

Products

• CA Automic Workload Automation - Automation Engine

Issue/Introduction

When monitoring an Azure Logic Apps workflow run, the job fails immediately even if a Failure Interval (e.g., 500 seconds) is configured. The job report shows the following error:

ERROR: Monitoring failed 
ERROR: Possible Reason: Request failed: 400 - Bad Request
Exception Response: {"Code":"BadRequest","Message":"Encountered an error (BadGateway) from host runtime." ... }

Azure logs may also show:

• Sc_status: 400
• Sc_substatus: 604
• RequestCanceled, ClientGone

Environment

• Integration.Azure.LogicApps 24.0.1
• Azure Logic Apps (Standard Tier) using ARR (Application Request Routing)

Cause

This was investigated under DE203206:

Azure Logic Apps uses ARR as a reverse proxy. When the backend host runtime becomes transiently unavailable, ARR terminates the connection and returns an HTTP 400 Bad Request with substatus 604 (ClientGone). While the 24.0.1 agent retries on 408 and 429 errors, it does not recognize this specific 400 error pattern as transient. The agent treats it as a permanent failure and stops immediately.

Resolution

Fixed in release Integration.Azure.LogicApps 24.0.2 and higher.

Additional Information

Solution details:

We have updated the integrations to systematically retry on all standard Azure transient HTTP status codes:

• 0 (Host not reachable)
• 408 (Request Timeout)
• 409 (Conflict - often a transient issue during ARM resource updates)
• 429 (Too Many Requests)
• 5xx Series (500 Internal Server Error, 502 Bad Gateway, 503 Service Unavailable, 504 Gateway Timeout)

Additionally, because Azure API Management occasionally wraps transient errors inside of a 400 Bad Request response, we have implemented specific parsing to detect and retry on transient messages within 400 responses, including:

• BadGateway
• ServiceUnavailable
• TooManyRequests
• TransientFailure