Skip to Content

Beyond the AIOps hype: Part 1

Sindhu Bhaskaran
2020-10-12

There is increasing adoption of AIOPS in the enterprise IT landscape. Gartner predicts that the use of AIOps and digital experience monitoring tools to monitor applications and infrastructure will rise from 5% in 2018 to 30% in 2023.

So, what is AIOps and why is it important for your organization?

AIOps, put simply, is the application of artificial intelligence, machine learning, deep learning, and big data to manage, automate, and improve IT operations.

Organizations are increasingly facing challenges due to:

  • Huge amounts of data provided by different monitoring systems. There is no way to get all the data and process it to get actionable information.
  • Many operations teams work in silos. Also, there is no way to visualize the system as a whole so there are likely to be delays in the diagnosis and resolution of issues.
  • Historical data analysis doesn’t help – there is a need to do real-time analysis of machine logs and data and provide insights for action or remediation.
  • Enterprises run into issues of critical outages, higher MTTR (mean time to resolve) issues, and reduced team productivity.

How do we enable AIOps to address these challenges?

To address these challenges, it is important for your organization to understand the key capabilities of AIOps and the processes involved in setting up end-to-end IT operations automation using AIOps. When we talk to customers, we recommend that the following five key processes be enabled for an end-to-end IT process automation using AIOps as depicted in the diagram below:

OBSERVE– Aggregate events/logs/alerts from all underlying systems (application, network, infra)   and enable real-time big data processing.

CONTEXTUALIZE – Map the underlying IT topology and enable end-to-end business service visibility across the different components of the business service or application. This includes capability for full-stack visualization – the ability to input or discover the topology is a key requirement.

THINK – Enable AI-based insights and recommendations using machine learning/deep learning. You could start with AI-based insights, which include noise reduction through event de-duplication and grouping, detecting anomalies in real time. Some of the advanced use cases could include event co-relation for causal analysis, automated RCA, prediction of application failures and change impact analysis.

ACT –Ability to initiate auto-heal/self-remediation workflows using RPA, ITPA, scripts, or orchestrators.

LEARN – Enable AI-based learning to learn from past events/failures and predict future scenarios.

Now that we know the different processes involved, we can follow the four key steps for a pragmatic implementation of AIOps for your organization.

  • Identify the right AIOps use case: Analyze the challenges and opportunities across the IT operations landscape to identify the use cases. You should make sure to include all the five processes above while defining the use case. You should make sure you identify the key KPIs you would like to impact – MTTD, MTTR, reduce ticket volumes, reduce outages or failures. We could start with some of the common use cases implemented for AIOps which include noise reduction, event co-relation, proactive detection of failures, automated root cause analysis, and change impact analysis.
  • Identify the right AIOps approach based on your requirements:
    1. Leverage best-of-breed big data, AI, and visualization stack for AIOps: This approach will enable you to start small with machine learning and build end-to-end automation use cases with AIOps. You could quickly demonstrate benefits without having to invest in new tools/vendors.
    2. Leverage AIOps vendors: They can provide out-of-box features and quick wins for noise reduction, event analytics, etc. Evaluate whether the tool can address all scenarios of end-to-end IT operations automation. Also, AIOps is an evolving journey and enterprises need visibility of how the data is gathered and how new insights and recommendations can be derived for end-to-end automation. This is a key requirement for scaling AIOps.
  • Start small with your AIOps journey: Create a technology-agnostic architecture, take an agile approach, and start small by gathering data, building artificial intelligence and machine learning models, gain insights and knowledge, and demonstrate end-to-end AIOPS use cases to deliver value. This will enable you to visualize and build the AIOps landscape incrementally and enable a state-of-the-art AIOps future for your organization.

To discover more about how CIAP provides a unique blend of technology agnostic, agile, and anti-lockin approach to assist you in creating enterprise wide AIOps at scale.

Author

Sindhu Bhaskaran