Comparative Performance Analysis of Integrated Monitoring Engine for Electric Energy Transaction Data Gateway Infrastructure to Accelerate SLA Incident Resolution

Abstract

PLN Icon Plus operates the Energy Transaction Data Gateway as the sole intermediary between banking partners and the national P2PST core server — an architecture where monitoring failure carries direct consequences for millions of daily transactions. Prior to this study, the monitoring ecosystem operated across three isolated platforms: Huawei iMaster NCE-Fabric for network telemetry, Zabbix for server resource metrics, and Elastic Stack (ELK) for application log management, with no automated correlation between them. This study developed an integrated monitoring system on the Grafana platform that unifies these heterogeneous data sources into a Single Pane of Glass dashboard. The architecture employs NTP-calibrated timestamp alignment and data normalization to ensure cross-platform event correlation accuracy at sub-100 millisecond precision. A unified alerting system was deployed via Telegram Bot API using multi-condition severity thresholding, requiring confirmed cross-layer correlation before notification dispatch to prevent alert fatigue. Comparative performance validation against the pre-implementation siloed condition — based on 69 documented production incidents from January to March 2026 — confirmed a 63.6% reduction in overall Mean Time to Repair (MTTR) and a 79.2% reduction in network incident MTTR specifically. SLA availability improved from 99.71% to 99.94%, surpassing the 99.9% contractual target. The primary contribution is a cross-layer data correlation model that measurably compresses the fault identification phase within national energy transaction infrastructure, validated through both statistical analysis and a structured questionnaire survey across 56 respondents.

Keywords

Full-Stack Observability; Data Gateway; Grafana; Mean Time to Repair (MTTR); Service Level Agreement (SLA)