Based on a recent industry study done by a major analyst firm, IT departments find that an excessive amount of time is lost when trying to determine the root cause of an issue. Often times this leads to dreaded MTTI (mean-time-t0-innocence) drills as multiple IT Teams are locked in the “War” room until the issue can be isolated and resolved. In those meetings there tends to be a lot of finger pointing (It Wasn’t Me – ft Shaggy), but what we need to ensure is that we don’t lose sight of the common goal to provide the best possible user experience. Is it a storage IOPs problem, a network latency, bad code, and how do you decide which tools to use to determine the cause of the issue?
A prevalent cause for extended periods of MTTI is a lack of collaboration among teams and management tools which leads to an inability to create context and dependencies through the IT infrastructure. Using vRealize Operations you can offset these shortcomings by understanding complex dependencies through autodiscovering relationships between IT infrastructure. This is possible by coupling vROps with endpoint management packs allowing endpoints to be tied directly to the virtual layer. Today we will build on previous blogs (Network, Compute, and Storage) to look at the entire IT stack from the database/application to the supporting infrastructure.
Figure 1 – Custom Relationship Dashboard
To reduce MTTI drills a single dashboard can provide the ability to quickly understanding the relationships through the stack and tie in any alert and KPI (key performance indicator) to understand how each layer impacts the next. Using the dashboard in figure 1 we can select any resource in the environment view (left side) and to the right the Health Status, Alert List, and Metric Picker will appear and provide deep insight around the selected resource. In a simple view it is possible to map a database or application into the virtual infrastructure and into the underlying hardware infrastructure. As an example it is possible to map an Oracle DB to the volumes of storage associated with it and through the entire stack in between using the specific management packs from Blue Medora at each layer. This allows a reduction in MTTI (mean time to innocence) by understanding the dependencies through the IT stack and seeing the key performance metrics and alerts at each level.
Figure 2 – Custom KPI Dashboard
Taking dashboards a step further we have created a custom KPI (key performance indicator) dashboard. The purpose of this dashboard is to build the relationships and dependencies via a Listview and tie in KPIs to the right. Let’s look at a use for for this dashboard. Say we are seeing an excessive amount of latency on a database. Is it the database that is causing the issue, or could it be the virtual infrastructure, or the underlying hardware. In the case of database latency we could map the resources all they way down to the volume to see if the cause is inadequate disk performance at that storage array.
Using the methods explored today, we can see how vROps can streamline troubleshooting to determine the root cause of an issue to provide the best possible end user experience, reducing MTTI and optimizing your performance in the process.