If the credentials of a queue are invalid (or invalidated) the queue monitor tries to query SQS indefinitely. The monitor should be stopped until credentials have been updated.
Changing credentials for a configured queue currently has no effect until Jenkins is restarted, if the monitor for that queue is already running. This means it is currently not possible for a user to fix #5 without restarting Jenkins.
it would be great if this plugin could support the use of the DefaultAWSCredentialsProviderChain so that an instance profile could be used to gain access to the sqs queue instead of requiring a separate user w/ aws credentials to be created.
The plugin currently creates multiple monitors for a single queue (race condition). This means multiple jobs that listen to the same queue are not necessarily registered to the same monitor.
A message will only be received by a single monitor, so only jobs registered to that monitor have a chance to be triggered. The monitor that receives each message is not deterministic, so messages will seem to (randomly) fail to trigger jobs.
As a workaround each repository and job needs its own queue which increases cost and is more difficult to set up.
I successfully downloaded and installed the SCM SQS Plugin for Jenkins and rebooted but when I tried to access the configuration for the plugin: Jenkins > Manage Jenkins > Configure System, I couldn't find any fields related to this plugin at all.
I'm using an AWS EC2 instance: Amazon Linux AMI 2016.03.2 x86_64 HVM GP2. I installed Jenkins using the installation instructions for Red Hat distributions, see here, resulting in version 2.8 of Jenkins being installed.
Update the documentation to mention that all services must be created in the same region. Add info that CodeCommit is only available in one region right now, so all services (SNS, SQS) must be set up in that region.