Share

Barclays has reduced the number of incidents experienced by users of its customer and employee-facing applications by 30 percent after implementing AppDynamics application monitoring tools.

Speaking at the AppD Summit in London last week, Barclays spoke about the benefits it is already seeing in app performance thanks to AppDynamics' application performance management (APM) tools. Barclays has more than 750 applications in its estate powering more than 12 digital interactions per second from customers alone.

Jamie Norburn, head of estate management at Barclays, explained how once the tool was moved out of a pilot deployment and into production, supporting its mobile apps for example, the IT team began focusing on incident avoidance. "Can we start spotting trends, deviations, logs, code issue that will stop those issues occurring?" We did," he said.

Read next: AppDynamics CEO David Wadhwani explains the logic behind that $3.7 billion Cisco acquisition, and where the vendor goes next

Norburn added that Barclays wanted to "really change application support from reactive to proactive" with custom dashboards, daily checks for service health, configure tiers and flow maps. 

Also, by freeing up developers from firefighting issues, and by giving them better visibility into where the apps are falling down, Barclays has seen "code getting better and the eradication of issues from the production environment". This has helped to stop issues from occurring before they get "anywhere near production" systems, Norburn said, adding that developers are also freed up to bring more value added feature releases for the apps.

Jim Korchak, managing director at Barclays, was unable to provide full metrics around the impact of deployment, but said during the question and answer session that customer incidents are down 30 percent since implementing AppDynamics and is seeing "significant improvements" in its time to recovery metrics.

It has also allowed the bank to start decommissioning legacy monitoring tools for a number of apps across the estate, but Norburn admits that "it's going to take some time at Barclays, simply because of the size of the organisation".

Read next: How Brexit helped Credit Suisse prove the value of AppDynamics' monitoring tools

Banks are especially sensitive to app incidents and downtime, as the recent outage of internet banking at Natwest, Lloyds and Halifax over the Mayday bank holiday would attest. Customers were unable to login and some saw money disappear from their accounts.

"Any person that is in the application support world and facing of to senior management is going to be worried about those things," Norburn said.

Read next: Bank IT meltdowns: Bank of England must do more to prevent bank IT failures, says Treasury Committee chair Andrew Tyrie

He added that simply implementing best-in-class monitoring isn't a silver bullet for issues, but that it helps the bank respond faster. "Incidents are going to happen," he said. "That is the nature of it, maybe once we have got this through our entire landscape I would have a bit more confidence. The point about APM isn't just eradication of incidents, it is about the insights it give you so the ability to land and remediate and get the service back online faster."

AI monitoring

Next Barclays wants to start using artificial intelligence (AI) and machine learning techniques to drive more automated monitoring and response across its app estate.

Read next: How technology will transform banking in 2017: Blockchain, cloud computing and digital challenger banks

Korchak said during the keynote: "Once we were able to see the wealth of data coming through AppDynamics a lightbulb went off that the amount of information might be too much for us to consume.

"So over the next three to five years we are looking to filter all of that useful information through artificial intelligence and automation so that the system will sense that something is going on and self-correct. If we can do that that is a game changer for our digital proposition.

"We want to understand and filter up that user experience so when it starts going south we can use that AI to understand what happened from an alerting finger print, and then we can start tracking how quickly we can automate or make that recovery."

Find your next job with computerworld UK jobs