Connector timeouts with ThingWorx Platform
If the eMessage Connector sends an AlwaysOn request to the ThingWorx Platform and the request timeout is reached before a response is returned by the ThingWorx Platform, the Connector will log this exception:
Connector Exception 21: TimeoutException: Timed out waiting from response.
Administrator Action
The administrator should take the following actions:
• Determine if the customer is using PostGreSQL for their ThingWorx platform persistence provider.
◦ If yes, ensure that the PostGreSQL server is running on a different computer than ThingWorx Platform. The PostGreSQL server should be installed on a different machine than ThingWorx Platform to ensure that no disk I/O contention occurs between database commits and the ThingWorx SystemRepository file system writes.
• Determine if timeouts occur when uploading files from agents.
◦ Review the logs and validate if the timeouts occur when the Connector is uploading file chunks to the ThingWorx Platform.
◦ Validate the write performance of the hard disk that contains the SystemRepository directory of the ThingWorx Platform.
▪ If using an Amazon EC2 EBS “General Purpose SDD” volume, consider migrating to a “Throughput Optimized HDD” or “Provisioned IOPS SDD” volume.
▪ ThingWorx has experienced extremely slow write performance on the General Purpose SDD volume under certain load conditions (13 seconds to write a 1MB file, for example).
• Increase the timeout for the ThingWorx Platform to respond to the requests from the Connector. The default value is 10 seconds (in the configuration file, emessage.conf). Note that this value is listed in milliseconds. The following example sets this timeout to 30 seconds (30,000 ms):
transport.websosckets {
request.default_timeout-ms = 30000
}