APM Terminals introduces industry-leading application monitoring solution

APM Terminals has commenced the global role out of an advanced application monitoring solution, capable of spotting issues with performance before they start to negatively impact operations. 

This new solution, which has already been rolled out to 11 terminals, will ensure business continuity, improve terminal efficiency, and safeguard consistency (peak moves per hour) by ensuring that applications consistently operate at the speed they’re supposed to. It will also prevent minor problems from combining to cause a major outage. 

According to APM Terminals Global Capability Manager, David Pickup:

“This industry first provides the same level of service normally only provided by advanced IT companies to their customers.” 

Traditionally the company relied on the basic monitoring offered by the suppliers of the company’s global standard terminal operating system (TOS) Navis. This year, the company’s TOS Support teams and Global Core Capability Centre (GC3) have established a new application monitoring capability that combines Riverbed and Elastic technologies. GC3 was established in 2018 to provide unrivalled internal technical capability around TOS.

Instead of simply monitoring isolated points, the solution monitors end-to-end performance – including code, application architecture, servers, disk space, databases, user endpoints and more. The data is fed into one central dashboard, which is monitored 24/7 by the Maersk global Command & Control Center in the UK and the GC3 support team. 

David Pickup says:

“The dashboard operates using a simple traffic light system. Green means that no issues are detected. Amber flags have the potential to impact the business.  Our goal is to fix these before they turn to red flags. Red flags indicate that the issue is probably already impacting our operations. This new solution enables us to identify any issues before the end user even notices that something isn’t working as it should.” 

Asked about real-life examples, Mr Pickup explains that the company’s databases should automatically purge themselves of old data, to ensure that they don’t run out of disk space:

“There are a number of reasons why this could go wrong. For example, a user running a query that blocks the database. At a specific threshold, we’d already want to be flagging this and investigating the reason before it impacts system performance.”

In some instances, fixed parameters, such as 96% usage of a database can lead to false alarms, so the team have applied machine learning to improve accuracy.

Mr Pickup explains:

“For example, say a terminal normally receives around 350 EDI messages per day from customers checking containers. A simple operating parameter might be between 150-550 messages before the system flags a potential issue. But in this example unexpected terminal closures on weekends or national holidays, seasonal fluctuations, or differences between different sized terminals could result in a number of false alarms. Instead of having fixed parameters, machine learning adapts to these to become more accurate and reliable in identifying real issues.”