Making Hybrid Infrastructure Monitoring Work for You

Most of today’s enterprises have significant gaps and overlaps in visibility across the datacenter, cloud, hybrid cloud, and containers, resulting in increased operational costs and the uncertainty of conflicting sources of truth. You need to see across your entire IT ecosystem (what you have, what it is, and how it works together), and understand how each component is related to each other (topology).

Organizations benefit from monitoring solutions that go beyond the infrastructure and show relationships through app-to-infrastructure mapping so that you can understand how the infrastructure and apps work together for a particular technology. Every application—from a complex financial services application to something as common as email— has a lot of moving parts. For example, if performance is slow, dependency mapping lets you know where to look to understand where the bottleneck is, what resources might be overtaxed, and how to resolve the problem.

Once you have these insights you know your next course of action should be to move workloads around or add capacity to handle spikes in usage.

Also, applications today are not running on dedicated infrastructure. They are relying on different components that are spread out throughout your entire IT ecosystem and can be shared with other applications. This makes having visibility across all your applications so important because it enables you to pull resources from other components to support others.

An example of this would be an eCommerce website needing to allocate more IT resources on Black Friday and Cyber Monday.

In a rapidly changing IT world with microservice-based, containerized workloads, monitoring at the device level is no longer practical. Forward-looking enterprises are shifting from tracking individual IT devices and apps to a single business services view across a heterogeneous mix of clouds.

In order to quickly resolve and remediate issues unique to your business, you must first understand your business from an IT perspective.

How do you do this? Begin by defining your business services.

Because understanding which services are critical and which SLAs you are expected to meet helps you know what you need to work on first in order to remediate and eventually automate.

What is a business service?

A business service is one or more technical services that gives value to both internal and external customers. Business services should align IT assets with the needs of a company’s employee, and customers and they should support business goals— facilitating the ability of the company to be profitable. Usually, a business service includes an associated Service Level Agreement (SLA) that specifies the terms of the service.

Examples of business services include:

• Verification of internet access or website hosting
• Remote backups and remote storage
• Payroll, online trading, online banking

There are many advantages to seeing across your entire hybrid IT ecosystem with a business service view. First, your IT devices are constantly in flux—from the software updates for your applications to the number of devices using your network at any given time. But your business services are a lot more stable and long-term—changing very rarely when compared to the dynamic complexities of IT components.

Having a long-term, service-centric view across your data centers, clouds, applications, and devices, helps you understand the impact of IT on the business. A service-centric approach to IT combines different applications together into a service your business is selling to the end user. And this way you can start managing/prioritizing your work— giving you the stability to understand and learn the behavior over time.

Another advantage of having a unified, service-centric view across your IT estate is having the ability to pull in business KPIs to measure ITs impact on business outcomes.

Why is this important?

Let’s use online trading as an example. A financial institution wants to achieve a certain number of trades a day. If they are monitoring their IT at the server-or application-level—the business is unable to see the high-level impact of those trades on revenue.

A service-centric approach takes away all the minutia and noise—giving you a clear understanding of how your infrastructure is impacting the service, so you can rapidly remediate before the end-user is impacted. If/when something goes wrong anywhere in your infrastructure—which is often shared across multiple applications and business services—you can quickly identify the impact and isolate the root cause.

By shifting from device-to service-centric monitoring, you can prioritize work based on business impact.

While infrastructure monitoring is essential in order to know what you have and what it is (discovery), and accurately assess the health, availability, and risk across your complex ecosystem, forward-looking companies are now looking to automate—eliminating manual operations in order to respond proactively and reactively. Now that you have composed business services and know the most common issues affecting your most critical services, you are ready to harness the power of machine learning to gain actionable insights from your infrastructure data.

Machine learning can detect weird or anomalous service behavior and correlate those anomalies and common events within a service context. With machine learning, you can cut through the noise to quickly establish the root cause of an issue—enabling your team to keep in front of constantly changing environments.

Adding machine learning to an infrastructure monitoring tool can unlock powerful opportunities for the ITOps team.

Developing learning patterns of behavior on the service and not the individual components gives you the ability to automate the right things. What are the “right things” you should automate? Let’s start with automated ticketing and routing. Industry averages indicate manual ticket creation and routing routinely takes more than 60 minutes. You can avoid costly service impact and downtime by eliminating these time-consuming, manual activities.

You can also automate troubleshooting and remediation steps.

By automating entire workflows, such as the steps required to remotely login to a device, gather diagnostics, and diagnose or remediate a problem, you can automatically capture diagnostic data to enrich both events and incidents.

This enables faster root cause analysis and improved mean time to repair (MTTR).

If you are automating data collection, you can improve problem management, then you will be better informed about what you can resolve. And once you know the commonality in that data—through mapping and monitoring by the service (not the device)—you can automate self healing and resolution of issues. When repetitive tasks and processes are automated, ITOps teams obtain the bandwidth to do the kinds of things machines are traditionally unable to do, including creative problem solving, upgrading existing technologies, and planning for the future.