Cisco IT’s observability technique

Within the face of rising IT complexity, Cisco IT unified observability throughout its international setting. The outcomes: 25% fewer main incidents, 45% quicker decision, and scalable automation. Right here’s how we did it—and suggestions in your digital resilience journey.

Our problem: fragmented visibility

Many organizations wrestle with fragmented monitoring and extended incident decision. We confronted the identical challenges and wanted to interrupt down knowledge silos to guard our fast-changing setting.

In Cisco IT, managing our international, dispersed IT panorama is more and more complicated. Fragmented knowledge and visibility gaps have been making it more and more tough to keep up digital resilience amid our fast-paced innovation and frequent environmental adjustments. We would have liked a method to flip our knowledge into actionable insights however lacked a unified observability platform to centralize and make sense of all of it.

When a significant database outage in 2024 revealed our fragmented knowledge — uncorrelated alerts throughout associated gadgets delaying trigger identification — we knew we needed to rethink our observability method.

“After the outage, we knew we needed to rethink every part. The transformation wasn’t nearly expertise, however empowering our engineers to behave quicker and smarter.”
– Chuck Churchill, Snr. Director, IT Observability, Cisco IT

This inflection level sparked a broader shift in how Cisco IT approached observability. Within the video beneath, Cisco IT leaders share what modified — and the way it led to a 25% discount in main incidents.

Getting began

Like every IT concern, understanding the foundation trigger was important earlier than remediation planning might start. Constructing stronger digital resilience is not any completely different.

In working with our groups, we pinpointed the next as prime problem areas that have been hindering our efforts:

• Fragmented knowledge and visibility gaps: With over 100,000 endpoints, we generated large volumes of telemetry knowledge. Siloed monitoring instruments led to alert fatigue and visibility gaps, slowing response occasions and limiting our means to foretell and stop points.
• Dangers from frequent adjustments: Our tradition of innovation and early adoption means fixed adjustments to our IT environments. We noticed a direct hyperlink between fast change and elevated incidents, so minimizing disruption with out slowing innovation grew to become important.
• Useful resource optimization: As our surroundings and knowledge complexity grew, it grew to become vital to enhance how we leverage AI and knowledge extra successfully. We would have liked to show our knowledge into actionable insights that empower engineers, not overwhelm them, to make sure productiveness stored tempo with development.

“Bringing all our knowledge collectively was step one—turning it into actual, actionable insights is what really allows us to remain resilient as our surroundings evolves.”
– Jon Heaton, Director, Community Engineering and Operations, Cisco IT

Deciding on our three-pronged observability method

Digital resilience requires greater than merely deploying new instruments. We would have liked a holistic method that might span our complete IT panorama. We obtaind this by structuring our IT observability follow throughout three interconnected pillars:

1. The community: Safe, dependable community efficiency is important for protecting the enterprise up and operating. This pillar focuses on complete community visibility, together with third–get together supplier networks, to make sure the community is operating securely and optimally for a easy consumer expertise.
2. Platforms and knowledge: Right here, we concentrate on observability throughout our knowledge facilities, cloud, and underlying infrastructure, centralizing our observability knowledge to be accessible to all the group, together with our DevOps and SRE groups, by means of our platform consolidation and knowledge technique.
3. Service Operations: Our Service Operations and Enterprise Operations Heart engineers monitor, analyze, and resolve points utilizing wealthy knowledge and insights delivered from our community, infrastructure, functions and providers — which feed AI automation to enhance effectivity.

Including within the vital expertise

Every of those observability pillars is powered by a mix of information, expertise, and processes that allow us to make use of the complete potential of our knowledge and drive higher effectivity by means of AI automation. These core components are important to our observability method and success:

• Splunk: Splunk acts because the spine of our observability technique, centralizing knowledge from throughout our community, infrastructure, and functions to ship a single supply of reality for our IT groups.
• ThousandEyes: ThousandEyes delivers end-to-end community visibility and consumer expertise monitoring throughout inner and exterior environments, enabling fast identification and determination of connectivity points.
• Configuration Administration Database (CMDB): Our CMDB gives a single supply of reality for all IT belongings, enriching alerts and incidents with important context and powering environment friendly, proactive operations.
• AI Operations: Our AI methods leverage the centralized observability knowledge in Splunk to automate occasion evaluation, cut back alert fatigue, and speed up incident response — empowering engineers to concentrate on higher-value work. For instance, our Service Operations makes use of numerous AI assistants for AI-driven incident administration to foretell incident assignments and counsel decision steps.

By integrating Splunk with ThousandEyes, our CMDB, and different instruments, we will guarantee a seamless, scalable method to observability that grows with our enterprise.

Tangible outcomes

This unified observability method has helped us sort out our most urgent challenges and enhance outcomes that collectively strengthen our digital resilience. Prior to now 18 months, we’ve seen:

• Vital discount in main incidents: We lowered main incidents by 25% 12 months over 12 months and had zero main community incidents, down from 3-4 per quarter beforehand.
• Sooner, more practical restoration: We lowered our Imply Time to Detect and Resolve by 45% 12 months over 12 months, enabling quicker restoration and minimized disruption.
• Improved change administration: We decreased incidents attributable to change by 20%, enabled by unified knowledge insights and end-to-end visibility that helps smarter change administration processes.
• Strengthened visibility and knowledge utilization: We now monitor 10x extra community telemetry knowledge that yields deeper insights and 4x higher visibility — enabling earlier detection and proactive decision of potential points earlier than they escalate.
• Scalable automation and effectivity:  Centralizing knowledge in Splunk has established a basis for continued developments in AIOps, enabling us to develop AI-automations that now deal with 99.998% of ~4 million each day alert — considerably enhancing operational effectivity.

“This journey is as a lot about altering mindsets as deploying expertise. We’re empowering each engineer to behave on insights, not simply alerts.”
– Mark Hutchins, Director, IT Service Administration, Cisco IT

Sensible takeaways

Our digital resilience journey is ongoing, however we be taught extra every day that we share with clients enhancing their very own journey. We suggest:

• Accumulate telemetry from in all places: Centralize telemetry to optimize essentially the most related metrics, occasions, logs and traces from throughout your community, infrastructure, cloud, functions, and providers.
• Prioritize knowledge – the bedrock of success: concentrate on knowledge high quality and hygiene. Correct, clear knowledge is important for producing reliable insights and enabling efficient automation.
• Use what’s already sensible: First, leverage the methods that you’ve got: product capabilities which have AI and observability inbuilt. Second, empower groups to experiment with customized AI options to maximise worth the place your current methods have gaps.

Keep tuned for updates and learnings as we proceed to innovate and optimize.

“Digital resilience is a transferring goal. We’re all the time studying, adapting, and refining our method as our surroundings evolves.”
– Jon Heaton, Director, Community Engineering and Operations, Cisco IT

Extra assets:

Discover extra case research to see how Cisco strengthens digital resilience