Designing a Unified IT Operations Monitoring Indicator System for Banks

Based on the information system lifecycle theory, a bank's information system can be divided into planning & design, development & testing (or purchase), implementation, and operation & maintenance with continuous improvement. The first three phases occupy only about 20% of the total cycle, while the remaining time is spent on running and maintaining the system, highlighting the critical role of IT operations services in banking IT construction and operation.

Background and Significance of the IT Operations Monitoring Indicator System

Business Need : To fully exploit the early‑warning function of operations monitoring, a systematic, standardized, business‑service‑oriented indicator system is required. It helps managers focus on decision‑making rather than complex IT details, presents information via dashboards, meets service‑level requirements, and supports overall business and IT decisions.

Regulatory Guidance : The China Banking Regulatory Commission’s "Guidelines for Data Center Supervision" mandates real‑time monitoring of critical information systems and networks, requiring monitoring records that support fault location, diagnosis, and post‑incident audit. This drives the urgent need for a practical monitoring indicator system.

Research on the IT Unified Operations Monitoring Indicator System

The system is a business‑service‑oriented, hierarchical, quantifiable set of monitoring indicators built on ITIL theory, linking IT management with business services. It aligns IT performance with business outcomes, improves customer satisfaction, and integrates ITIL processes with Business Process Management (BPM) to provide a visual, one‑stop view of IT service status.

During design, the Analytic Hierarchy Process (AHP) is used to decompose monitoring elements (network, host, middleware, database, application, etc.) into interrelated units. Experts evaluate the relative importance of each unit, producing a relationship matrix and calculating weights. The final hierarchy consists of four layers: Application Service, System Resources, Network Services, and Infrastructure, covering all domains such as applications, databases, middleware, servers, storage, network, and power environments.

Indicator categories include:

Application Service Layer : transaction process, transaction data, batch processing, transaction logs/messages, error information.

System Resources Layer : database (server status, instance status, sessions, locks, listeners), middleware (WAS, WebLogic, MQ), operating systems (Windows, Linux, Unix), storage (fiber switches, ports, systems, links).

Network Layer : device CPU, memory, fan, temperature, power, system, ports, protocols, etc.

Infrastructure Layer : power meters, UPS, air‑conditioning, etc.

Standardized data collection interfaces aggregate, classify, and correlate these metrics, enabling event, performance, alarm, and fault management, as well as daily supervision of batch jobs, backups, version control, maintenance, on‑call duties, and asset management (see Figure 1).

Construction Methodology

The process follows five stages to create a sustainable monitoring loop: indicator identification, indicator setting, weight calculation, evaluation, and system establishment.

1. Indicator Identification

Development and operations teams gather business characteristics and system contexts, using expert surveys to select a core set of indicators (see Table 1). Images illustrate the collected indicator set.

Performance indicators use three threshold levels: baseline (normal operation), attention (optimization trigger), and alarm (capacity‑expansion trigger). The principle baseline < attention < alarm guides threshold setting, which can be refined based on operational data.

2. Indicator Setting

Teams define collection method, frequency, data type, alarm conditions, severity, description, output interface, and field format for each indicator. Critical indicators are marked (e.g., with an asterisk) and have explicit thresholds, as demonstrated in Table 2.

3. Weight Calculation

Weighting can be performed quantitatively with AHP, which offers reliable pairwise comparisons, or qualitatively with expert judgment, which is simpler but may be less accurate when many indicators are involved.

4. Indicator Evaluation

Indicators are assessed using the SMART criteria: Specific, Measurable, Attainable, Relevant, and Time‑bound, ensuring each metric is actionable and aligned with operational capabilities.

5. System Establishment

After clarifying which indicators to monitor, their hierarchy, dependencies, and importance, a layered indicator system is constructed (see Figure 2).

Conclusion

Establishing a unified IT operations monitoring indicator system is fundamental to effective ITIL implementation and risk mitigation. It enables quantitative analysis combined with qualitative judgment, integrates development and maintenance perspectives, and provides timely alerts and comprehensive analysis to both prevent operational risks and optimize application performance.