Latest posts by Paul Schnackenburg (see all)
- Azure Data Lake overview - Fri, Sep 22 2017
- Moving from Office 365 to on-premises Exchange - Tue, Sep 19 2017
- Office 365 Secure Score – Securing Exchange Online - Thu, Aug 3 2017
With the extremely long and public gestation period (the first technical preview was in late 2014) of the next version of System Center (SC), it’s interesting to follow the evolution of each component of SC. Now in Technical Preview 4, Operations Manager is going to bring the ability to monitor Nano server (as well as full Windows Server 2016), scheduling of maintenance mode, full LAMP stack monitoring, and management pack recommendations in the console.
For an in-depth primer on Operations Manager, look at our series on the 2012 version here, along with the update for the 2012 R2 version here. Bear in mind that the RTM version of Operations Manager 2012 R2 has been updated with frequent Update Rollups (as has the rest of SC), bringing additional functionality.
Keeping with the trend at Microsoft to embrace the open source software world, Operations Manager 2016 will offer full monitoring of the LAMP stack. Back in 2012, Operations Manager introduced monitoring of Java application servers, and as of Update Rollup 4 for 2012 R2, the following JEE platforms are supported:
- Tomcat 5.5, 6.7, and 8
- JBoss 4.2, 5.1, 6, 7, and 8
- WebSphere 6.1, 7.0, and 8.x
- Weblogic 10g Rel3, 11g Rel1, and 12 Rel 1
There’s even in-depth monitoring using an open sourced component called BeanSpy that gathers MBean counters (sort of like performance counters in Windows); for a detailed look, see here.
But the Java platforms is only a small part of the puzzle. Linux, of course, has been supported with native Management Packs (MPs) since the Operations Manager 2007 days, and today many *nix flavors are supported, including RHEL 5-7, Ubuntu 14.04, and SLES 12, along with many Unix versions. Missing has been Apache and MySQL support, but this is coming in Operations Manager 2016 in the form of new MPs building on the Operations Manager Linux agent and Open Management Infrastructure (OMI).
Both MariaDB and MySQL servers and databases are dynamically discovered, and performance metrics for key, query, and table caches are gathered. Optionally, you can collect metrics for the InnoDB buffer pool, and you can also extend the monitoring with stored procedures. You do need credentials for the MySQL instances, but it can be a fairly low-privilege account; it only needs permissions to read the performance data.
The Apache MP relies on an Apache module that is installed automatically if the Linux agent discovers that Apache is installed. It monitors the Apache server itself, virtual hosts, and SSL certificates. You can optionally create probe rules and monitors for availability and response time information. Of immediate value is the 21-day warning and 7-day error alert on SSL certificate expiry. There’s also a handy built-in task to reload the Apache server. It also lists all modules that are installed on each Apache server.
Another notable improvement is the ability to run a full shell script or Python/Perl script to identify the health of custom applications. In the current version, command line monitoring for Linux is limited to a single command line; it’s also downloaded to the client as an XML file, which means that angle brackets can cause issues. The script engine in Operations Manager 2016 is multithreaded for parallel execution; the current single-threaded engine can queue up items behind a long running script.
Credentials for *nix machines can now be stored as Run As accounts so that operations staff can use these to discover servers without having to know the password. There are many different file systems in use on Linux; some of these are virtual, and you can optionally filter out particular file systems during discovery.
In part 2, we’ll look at other improvements in Operations Manager 2016, such as scheduled maintenance mode, Storage Spaces Direct, and Nano Server monitoring, as well as the new MP catalog.