This large organization was using 6 different monitoring products. SCOM for windows, BMC for Linux, Riverbed Steel Central for APM, Oracle OEM for Oracle Databases, SCOM for SQL Server databases, VMWare for VM monitoring, and Empirix for web transaction monitoring. Some of these were deployed across the enterprise and others were deployed departmentally. The goal of the Foglight implementation was to consolidate vendors while improving the value being delivered to the benefiting teams, and centralizing the system monitoring function. This was no small task given that the implementation spanned 34,000 physical and virtual operating system hosts across 14 data centers.
The most pressing need was to establish a standard for the monitoring of operating systems. This meant doing what SCOM was doing for Windows, what BMC was doing for Linux operating systems, and what VMWare was doing for monitoring of their own virtual infrastructure. It was important that we didn’t just meet the basic requirements. We had to go above and beyond because the current performance monitoring solutions were not providing adequate value and many were considered to be dated.
In addition to replacing the operating system monitoring, the team was also very interested in the monitoring of SQL Server databases. The customer was using SCOM to perform this function but it was falling short on meeting their needs. In initial demonstrations, the DBA team was adequately satisfied with what Foglight’s SQL PI provided and wanted to see that implemented as well. Foglight's SQL PI was considered to be well above and beyond what they were using prior. There were approximately 1500 SQL Server databases in scope.
Lastly was Synthetic Web Transaction Monitoring. Empirix was being used to validate the performance and availability of web servers, a function that is in Foglight’s wheelhouse. After demonstrating Foglight’s web monitoring functionality the product came up short in 2 areas; localization (l10n), and error checking of content. Besides those 2 short comings Foglight was a suitable replacement for Empirix.
Current Status
This holiday season was a major milestone for the Foglight project. It was fully deployed for all production windows hosts. Foglight was performing basic host resource monitoring and log file monitoring, in place of SCOM for over 14,000 hosts. Additionally, the Foglight R&D team was able to close two outstanding enhancement requests, where we had fallen short on replacing Empirix. As a result we did get Foglight’s Web Monitor installed to serve as the primary for synthetic validation. Lastly, we did have some of the SQL Server monitoring in place for the first time. Much of the DBA team was able to use it for performance diagnostics.
Work is not complete but the team is moving forward with next steps. Customization that supports monitoring provisioning and de-provisioning automation are in progress along with the deployment of Linux host monitoring. There is also continued work to be done on the deployment of monitoring all SQL server databases.
Widening Scope
Besides what’s being monitored there are plans to review some of the Microsoft application monitoring features of Foglight. There are also road map reviews scheduled to talk about many of the new Foglight features being added to further support large infrastructures such as this. Lastly, the node count has nearly doubled since we started as a result of acquisition.
Summary
In the beginning there were 6 monitoring product and now there are 3; Foglight, OEM for Oracle, and APM (TBD). So far Foglight has replaced or nearly replaced all of the following:
- From SCOM for Windows Host Monitoring -> to -> Fogilght for Windows Host Monitoring
- From SCOM for SQL Server Monitoring -> to -> Foglight SQL Performance Investigator (PI)
- From BMC for Linux Host Monitoring -> to -> Foglight for Linux Monitoring
- From Empirix Synthetic Web Monitoring -> to - > Foglight Synthetic Web Monitoring
- From VMWare for VMWare Monitoring -> to -> Foglight for VMWare Monitoring
An ROI for the reduction of man power, reduction of monitoring infrastructure, and reduction of licensed monitoring products has not been performed. There are however, many data point that speak to project's initial success. The teams are finding more problems earlier, there are now 3 less monitoring interfaces to learn and login to, and last count we showed 380 active Foglight users.
More to come in 2017!!
