Background:

Multinational banks are at the forefront of digital transformation, as they deal with shifts in customer expectations, regulation and cost. In a nutshell, a banking application cannot afford slowness or display even the slightest hint of downtime. This places a considerable amount of importance on transaction turnaround time or average end-user response time.

READ MORE

What is transaction turnaround time?

Transaction turnaround time is a metric that paints a precise picture in relation to how many seconds it takes for an application to receive and process an end-user’s request; be it the start to end processing of a business transaction or even a swipe, tap or a click.

In short, it is a quick indicator of the digital customer experience.

Problem:

The multinational bank experienced intermittent slowness initially, before spiraling into all-out sluggishness. Here’s a breakdown of the event timeline, detailing how AppsOne & OpsOne combined to fix the problem.

Event Timeline:

Problem:

  1. 6:52 pm: First indication of premium banking intermittent user experience slowness.
  2. 6:53 – 7:37 pm: Generally good user experience with brief, intermittent, moderate slowness.
  3. 7:38 pm – 7:39 pm: Major slowness experienced by users for two minutes.

Solution:

  1. 7:39 pm: Appnomic AppsOne kicks off an automation through OpsOne to turn down a secondary service and free up resources for user demand and experience.
  1. 7:40 pm: User experience back to normal and continues at even better performance – crisis avoided.

Value Proposition:

  • Improved response times
  • Stickiness
  • Premium digital customer experience

READ LESS

Preventive Analytics-Based Alerts:

Background:

Auto Discovery is a feature that dynamically discovers and adjusts performance thresholds. Pattern matching techniques are used to generate early warning alerts to isolate fault and provide actionable insights; thereby eliminating the manual efforts required to set static thresholds and correlate KPI data across tiers.

What is an Early Warning Alert™?

To understand how our Early Warning Alert works, we must know the difference between static and dynamic thresholds.

READ MORE

A static threshold is a predefined parameter. For example, if we predefine a condition stating that an alert should be generated if and only when the CPU utilization exceeds 4%, it’s a static threshold. From a business context, a static threshold is flawed since it ignores real-time changes in relation to volumes and traffic driven by end-users. On the other hand, dynamic alerts are driven by a proactive monitoring and preventive analytics technology. Our Early Warning Alert™ is one such analytics-based alert.

A dynamic threshold is not predefined. When an analytics software is installed in an IT environment, it performs a statistical analysis on all the data points across all infrastructure layers to discover patterns and establish baselines. And this analysis leads to the detection of performance anomalies. The preventive analytics technology compares the anomaly to what was earlier identified as ‘good behavior’ and discovers dynamic thresholds in real-time as and when the deviations occur. This is where the Early Warning Alertsteps in. Post discovery of the dynamic threshold, we can configure the analytics-engine to automate the problem remediation process to trigger alerts and remedial actions.

Problem:

ABC Bank is growing at the rate of knots. This growth has direct repercussions on the number of transactions the bank must process on a given day. Of course, this spike in volume will eventually snowball into ‘downtime’; compromising the bank’s promise to deliver and sustain high-quality digital customer experiences.

Event Timeline:

Problem:

  1. 9:30 am – 10:00 am:  CPU utilization rose from 5% to 15% utilization within a half hour.
  2. AppsOne generated an Early Warning Alert™ on the abnormal behavior of a database CPU.
  3. AppsOne alert was left unattended due to disbelief.   No other systems monitoring alerts were generated by other tools.
  4. 4:00 pm – CPU hovered around 15% until just after 4:00 pm – 6 hours.
  5. 4:30 pm – the database stopped responding and Internet banking application users experienced an outage.

Solution:

ABC Bank had a visceral experience about how AppsOne Early Warning Alerts™ works and they now use AppsOne as a core part of their systems management and analytics environment.

Value Proposition:

  • Uptime
  • Premium digital customer experience

READ LESS