In this next instalment of our Designing for the Cloud series, we look at how to make management of a functionally rich cloud contact center platform... well, manageable.

For the landlord of such a system (usually a service provider), when things go wrong, 99 times out of 100 the issue is related to environment, e.g. carrier, network, database issues.

How do we know this? Because Sytel's support team has been providing technical support to datacenter-hosted and cloud-hosted service providers for almost 20 years now and have built up a pretty good understanding of the challenges service providers face when trying to troubleshoot end-user problems.

In order to find a resolution, someone has to come quickly to an understanding of the problem. Speed is of the essence. A customer might be forgiving of a preferred carrier outage if this is rectified in minutes by a temporary alternate route, but taking hours to resolve is not an option.

Coming to an understanding of the problem usually requires looking at log files from various sources. This would normally involve access to the operating system console, but in the case of a distributed system, locating the right logs on the right server can eat up precious time. And this approach comes with considerable security and fraud prevention concerns.

So, the key to reducing resolution time is to eradicate this need for console access for managing the platform day-to-day. Sytel achieves this by having a consistent scheme for logging for all services, and by funnelling web server, media server and database server logging into the same infrastructure, providing a single, web-based point of access to all log information.

For a landlord with a customer who has a specific agent with problems, this consolidated logging and log browser enables them to map activity across all services for a specific agent over a time period, and then expand to see context. And this extends to all resources within the contact center - media sessions, queues, campaigns, routing decisions.

Security is maintained by

  1. providing authentication-based access to log data using HTTP over SSL, and
  2. running the service for each tenant as well as a landlord instance to ensure scale and tenant data separation

Using this toolset, troubleshooting can be done exceptionally quickly, allowing the problem to be swiftly rectified, or asking the end user to take the necessary corrective action.