Remote Monitoring of Enterprise Systems A Step Towards Effec1ve Management of Cloud Based Services Johnson L Fisher, Director, IS Opera5ons May 28, 2015
Agenda Overview Current State Facility and Service Monitoring Enterprise System Monitoring History Transi5on Steps Impact on Services Conclusion Agenda
Overview Managing Data Centers and the services they support has become more complex due to the impact of virtualiza5on and the uncertain5es of what, when, and how much will be provisioned using cloud based services. The Monitoring of Cloud Based Services must be inherent part of the successful adop5on and Management of Cloud based services. Overview
Current State of Enterprise Systems Monitoring at KSU Data Center Facility Management: Data Center Facility Improvements Facility Monitoring & Management Tools Opera1ons Remote Data Center Monitoring: Updated Tools, Updated Procedures Opera5ons Control Center Implementa5on of Enterprise Monitoring Suite Enterprise Monitoring Suite: BMC Proac5veNet Performance Management CompuWare Data Center Real User Monitor New Relic End to End Transac5on Inspec5on Current State
Hot Aisle, Ceiling Air Return Data Center Facility Improvements CRAC Unit, Extension, Floor Air Supply
Opera1ons Control Center
Enterprise Monitoring History Complex Technology Stack Unsuccessful root cause analysis Time spent auemp5ng resolu5on Unnecessary Hardware Purchases Monitoring Tool Disconnect Oracle Grid SiteScan VMCenter Splunk Eagle- Eye Challenges
Enterprise Monitoring History - con1nued Process and Tool Deficiencies Challenges No event correla5on between tools No central Event Management Inconsistent Incident Management Silos Persist
Transi1on to Enterprise System Monitoring Opera1ons Remote Data Center Monitoring: Detailed Analysis of Data Center Ac5vity Iden5fied What Could Be Done Remotely Completed Required So]ware/Firmware Upgrades. Implementa5on of Improved Airflow Management IS Opera5ons Remote Procedures Manual Implemented Remote DC Monitoring Challenges
Transi1on to Enterprise System Monitoring Enterprise Monitoring System Implementa1on Ac5vity BMC Proac5veNet Performance Manager (BPPM) So]ware selected as our Single Pane of Glass Interviewed & Selected Consultant to Implement BPPM Procurement Tasks so]ware and hardware (including hardware design) Installa5on Hardware and so]ware Timeframe December 2013 January 2014 January February 2014 March April 2014 Project Summary Project Kick- off APM DC RUM & Synthe5c so]ware selected for end user monitoring Procurement Tasks so]ware and hardware (including hardware design) Installa5on Hardware and so]ware APM New Relic so]ware selected for end user monitoring Training BPPM & APM OCC So] Launch with BPPM Project Celebra5on March 2014 March 31, 2014 April March 2014 March April 2014 June 19, 2014 April October 2014 August 2014 October 31st 2014
Transi1on to Enterprise System Monitoring Enterprise Monitoring System Implementa1on 30+ Training Sessions 160+ Training Hours 744+ Consultant Hours 23 BPPM Specific Views with Graphs for Service Owners 17 Applica5ons monitored 50+ Transac5ons Monitored 56+ People Assigned 102 Patrol Agents Installed 2,400+ Events handled by OCC from BPPM 6700+ Project Hours Project Summary
Transi1on to Enterprise System Monitoring Enterprise Monitoring System Implementa1on Opera1ons Control Center: Opera5ons Remote Monitoring Invi5ng Space for Informa5on Sharing Scorecard Review Cross Training Sessions Crea5on of a Flexible Space War Room Event Management Problem Analysis Incident Management Problem Management Project Summary - OCC
Opera1ons Control Center
Transi1on to Enterprise System Monitoring Enterprise Monitoring System Implementa1on Opera1ons Control Center: BPPM Heat Map Scorecard Review Help Desk Ticket Tail Network Monitor Cri5cal and Major Events; All Events Cisco Wireless Monitor Data Center Security Cameras Weather (Web Based and Cable Feed) Cable Feed Guest Feed Project Summary - OCC
Transi1on to Enterprise System Monitoring Enterprise Monitoring System Implementa1on Impact on Services: Real User Monitoring Synthe1c Transac1on Monitoring Collec1on of Baseline Data Focus on Ac1onable Informa1on Project Results For On Premise and Cloud based services
Conclusions Does Lights out decrease outages? Theory says yes, however, we have no sta5s5cs to support this. We have reduced Data Center traffic and access to the Data Center is a much cherished privilege. All working in the Data Center must do more planning. Does Remote Monitoring with Enterprise Monitoring tools assist with managing cloud based services: We believe the answer here is clearly yes. We can detect through Real User monitoring. We can measure transac5on 5ming. We can build models of normal baselines and signal events from divergences. Project Results
Conclusions Does Enterprise Monitoring improve the end user experience. We believe the answer is clearly yes. We know we discover RUM issues faster. We don t have a baseline from the past. We can measure Outages however, that number may increase just because we more accurately discover outages. We have detected ini5a5ng events that lead to outages. Some of these can be repaired before they cascade into an outage. o We can automa5cally repair (e.g. stop and restart a service) before the outage occurs. Project Results
Conclusions Summary Remote Monitoring embodies a mindset, a toolset, procedures, and a commitment to a different opera5onal model. This opera5onal model has the flexibility and approaches to support the effec5ve monitoring of cloud based or hybrid cloud sourced services. The effec5ve monitoring of cloud services along with procedures and the administra5ve and technical resources offer the best opportunity to effec5vely deploy and managed Cloud based Services. Project Results
Conclusions Ques1ons? Contacts Questions? Johnson Fisher, Director, Informa5on Services Opera5ons jfishe16@kent.edu 330 672 1330 Enterprise Monitoring Core Team o Todd Ryan, Manager, Enterprise Monitoring o David Veits, Technical Lead o Amanda Kelley, System Administrator o Shelley Sherwin, Project Manager