[AMQ-8255] Wrong exception propagation on reaching Broker Connection Limit

• Type: Bug
• Status: Open
• Priority: Critical
• Resolution: Unresolved
• Affects Version/s: 5.15.13
• Component/s: Broker, JMS client

Environment

Broker Details:

• Version - 5.15.12
• Hosted on AWS (Amazon MQ)
• Instance type - mq.t3.micro with a limit of 100 connections

Client Details:

• Version - 5.15.13
• Using openwire/jms client
• running on linux **

• Assignee: Unassigned
• Reporter: Shashank Krishnaswamy
• Created: 30/Apr/21 21:49
• Updated: 30/Apr/21 21:49

Description

Hi,

I am trying to add exception handling at activemq client side around connection setup. Based on my experiment I found that whenever Broker reaches the maximum connection limit, client receives 3 types of generic exceptions - SSLHandshakeException, SocketException & TransportDisposedIOException. These exceptions could occur if there is connectivity issue or truststore/certificate related issue. This has made exception handling impossible as it is not possible to know if the error is retry-able (recoverable) or not.

The ask here is to provide better signal from Broker (or client library) so that application can handle recoverable and non recoverable errors differently.

Suggestion options

1. Let connection.start() throw ExceededMaximumConnectionsException
2. Let client provide exception via listeners
   • Client internal exception listener
   • Exception listener
   • Transport Listener
3. Or provide new API to get current number of client connections.

Thanks