Hello everyone,
I'm investigating a strange behavior in a Blue Prism environment and would appreciate any insights from people who may have seen something similar.
Environment
- Blue Prism 7.3.x
- Separate DEV and PROD environments
- Dedicated Application Server and database for each environment
- Runtime Resources connect through Application Servers (no direct database connections)
- WCF SOAP with Transport Encryption
- ASCR enabled (gRPC)
- Runtime Resources use the standard Resource PC listener on port 8181
Problem Description
One specific Runtime Resource behaves differently from all other Runtime Resources in the environment.
The resource is explicitly started using the PROD connection:
/resourcepc /public /dbconname PROD_Connection
`
The listener log shows:
checking connection 'PROD_Connection'
connected to 'PROD Application Server'
Listener started
The resource appears correctly in the PROD Control Room and functions normally:
- Processes can be started successfully
- Processes run and complete successfully
- The resource remains online in PROD
- No unexpected disconnects from PROD occur
From a production perspective, everything appears to work correctly.
Unexpected Behavior
Despite being connected and fully operational in PROD, the listener continuously receives incoming connections from the DEV Application Server.
Example:
New connection from DEV-AppServer
Disconnected ... quit
These connection attempts occur approximately every 30 seconds.
What is important here is that:
- The Runtime Resource is connected to PROD
- The DEV Application Server initiates the connection
- The Runtime Resource is not attempting to connect to DEV
Why This Seems Unusual
Other Runtime Resources in the environment:
- Use the same configuration structure
- Have the same connection definitions available locally
- Are known in both DEV and PROD environments
- Use the same startup mechanism
However, none of them exhibit this behavior.
The issue is isolated to a single Runtime Resource.
Investigation Already Performed
Runtime Resource Side
- Compared Automate.config with working Runtime Resources
- Compared LoginAgentService.config with working Runtime Resources
- Explicitly used /dbconname PROD_Connection
- Reinstalled Login Agent
- Rebooted the machine
- Confirmed only a single Automate.exe instance is running
- Verified listener functionality on port 8181
Database Investigation
- No active or hanging sessions found for this Runtime Resource
- No open sessions found in BPASession
- Historical schedules referencing the resource were found
- Schedule execution logs show no recent executions
Scheduler Test
To determine whether the DEV Scheduler was responsible for the connection attempts:
- Scheduler disabled on DEV
- DEV Application Server restarted
- Connection attempts continued
Therefore, the Scheduler appears not to be the cause.
Application Server Test
When the DEV Application Server service is stopped:
No incoming connections from DEV
`
When the DEV Application Server service is started again:
Incoming connections from DEV immediately resume
This strongly suggests that the DEV Application Server itself is actively initiating these connections.
Main Question
Has anyone experienced a situation where:
- A Runtime Resource is connected and operating normally in Environment A (PROD)
- A second Application Server from Environment B (DEV) repeatedly attempts to connect to that Runtime Resource
- The Runtime Resource remains fully functional in PROD
- Other Runtime Resources in the same setup do not show the same behavior
Additional Concern
At the moment this does not appear to be causing any functional issues because the Runtime Resource remains connected to PROD and can execute processes normally.
However, my concern is understanding why the DEV Application Server is attempting to communicate with this specific Runtime Resource every 30 seconds.
I am worried that at some point DEV could inadvertently interfere with the Runtime Resource, create session conflicts, affect scheduling behavior, or otherwise interact with a machine that should currently be operating exclusively in PROD.
Areas I'm Currently Considering
- ASCR (Application Server Controlled Resources)
- gRPC callback registrations
- Resource registration metadata
- Stale runtime references
- Internal health checks
- Callback channels
- Load balancing or Application Server routing behavior
Has anyone seen similar behavior or have any suggestions on where to investigate next?
Thanks in advance for any ideas or guidance.
Michael
