AIOps Boot Camp

Part 1: Introduction to AIOPS

We'll start with some specifics of what AIOps can do in terms of corporate management. 

1. Reduce management overhead
2. Reduce human error
3. Enhance compliance with policies and procedures
4. Increase accountability
5. Increase enterprise reliability
6. Improve response time
7. Decrease need for help desk and first-level engineering labor
8. Develop a holistic view of corporate data
9. Achieve an end-to-end view of corporate intelligence

Then there are benefits that are specific to IT: 

1. Reduce mean time to recovery
2. Simplify troubleshooting
3. Implement automated remediation of events
4. Automate routine processes
5. Implement alert filtering

Part 2: AIOps Drivers

The demand for digital transformation is driving AIOps. These are some of the issues AIOps addresses. 

1. Massive amount of data being produced, stored, and analyzed
2. Rapid pace of digital transformation
3. Marketplace demands
4. Review of Gartner AIOps Market needs report
5. Proliferation of Internet of Things devices
6. Cloud migration
7. Class Exercise: Identify complexity drivers in your organization. In this exercise, you'll examine your company's current situation and identify issues that AIOps addresses.

Part 3: AIOps Architecture

These are the building blocks of an AIOps platform. 

1. Data ingestion
2. AI and ML engines
3. Big data
4. Action through automation and enhanced analytics
5. Review Gartner's AIOps architectural model

Part 4: Machine Learning

Machine learning creates the intelligence that powers AIOps. The success of an AIOps initiative depends on the accuracy of your ML model. 

1. What is machine learning?
2. Descriptive statistics 
   1. Linear regression
   2. F(x) = y = mx + b
3. Supervised learning
4. Unsupervised learning
5. Reinforced learning
6. Data training
7. Feature vectors
8. Neural networks
9. Decision trees
10. Clustering
11. Model learning
12. Model verification
13. Model deployment
14. End user acceptance
15. Class Exercise: Build a TensorFlow Quickset with Keras that trains an ML model.
16. Class Exercise: Build, train, and deploy a machine learning model with Amazon SageMaker.

Part 5: Data

AIOps effectiveness depends on the quality and amount of data fed into the ML model. The old saying "garbage in, garbage out" applies directly to AIOps. 

1. Selecting data sources 
   1. Data selection is critical to model development because it provides the variables for an ML model.
   2. Structured sources
   3. Unstructured sources
   4. Contextual sources
2. Data mapping to AIOps configuration management database
3. Service mapping
4. Event collection 
   1. Filtering
   2. Deduplication
5. Agentless discovery
6. Class Exercise: Map and import event data using ServiceNow IntegrationHub.
7. Time series database 
   1. Class Exercise: Set up Timestream database on AWS.

Part 6: Enhanced Service Desk

AIOps improves the efficiency of your service desk through the use of the following tools. 

1. "Single pane of glass" for all monitoring systems 
   1. Class Exercise: Import a prebuilt Grafana dashboard into Prometheus.
2. Automatic monitoring
3. Automated incident response
4. Automated remediation
5. Intelligent service orchestration 
   1. Interactive automation
6. Alert correlation 
   1. Correlation versus rules
   2. Correlation model fields
   3. Correlation model input sources
   4. Moogsoft alert correlation example
7. Root cause analysis
8. Predictive analytics
9. Knowledge base
10. Natural language processing
11. Chatbots
12. Anomaly detection 
    1. Anomaly definition
    2. Adaptive thresholds
    3. Multivariate versus univariate
    4. Static threshold versus dynamic threshold
    5. Time series data 
       1. Stationary
       2. Seasonal
       3. Trend

Part 7: Automation and Service Orchestration

Automation of processes is a major benefit of AIOps. This section covers AIOps and automation. 

1. Challenges 
   1. Where do we find the time?
   2. What do we automate?
   3. Organizational silos
   4. Updating
2. Scripting for automation 
   1. Human-generated and human-developed
   2. AI-triggered workflows
   3. Open-source tool Ansible
   4. Class Exercise: Create automations with the ServiceNow workflows.
   5. Class Exercise: Integrate Ansible scripts into Moogsoft AI.
3. Intelligent automation—Automation Anywhere's discovery bot 
   1. Enables automatic process documentation
   2. Identifies correct process for automation
   3. Enables robotic process automation developers to accelerate automation
   4. Converts automation opportunities into bot prototypes

Part 8: AIOps Use Cases

Learn how AIOps helps IT achieve goals! 

1. Reduce mean time to recovery through automated remediation
2. Reduce mean time to failure through predictive analytics
3. Reduce delinquent payments through automated document processing
4. Reduce development time by automating the DevOps pipeline
5. Refocus IT resources on high-level initiatives by automating the level one help desk

Part 9: AIOps Implementation

This section explains the basics of getting started with AIOps. 

1. Top management support for the elimination of silos
2. Understand business plans
3. Present systems analysis 
   1. Involve all affected business units
   2. Detail and inventory current systems
   3. Identify problems 
      1. Quantify hard costs
      2. Identify soft costs and have business units agree on value
   4. Identify all sources of data
   5. Asses staff resources
4. Chose a use case
5. Class Exercise: Develop a business case for upper management to implement your AIOps use case.
6. Platform selection 
   1. Domain-centric
   2. Domain agnostic
   3. Open-source proprietary
   4. RFI/RFP development
   5. Cloud-based
   6. SaaS
   7. Hybrid
7. Creating a data lake
   1. Determine data sources
   2. Collect data into centralized AIOps database
   3. Analyze data
   4. Cure data
8. Apply AI engine to data to create use case model
9. Test model against historical data
10. Deploy model
11. Use AI to document model
12. Test model in real time
13. Integrate contextual data from end users
14. Enable AI and ML to integrate feedback into model

This interactive training course is designed for technology professionals looking to expand their knowledge of AIOps.

Professionals who may benefit include:

• IT Leadership
• System Administrators
• Developer Team Leads
• Software Managers
• IT Operations Staff
• Configuration Managers

You will learn how implement AIOps so that you can:

• Reduce management overhead
• Reduce human error
• Enhance compliance with policies and procedures
• Increase accountability
• Increase enterprise reliability
• Improve response time
• Decrease need for help desk and first-level engineering labor
• Develop a holistic view of corporate data
• Achieve an end-to-end view of corporate intelligence
• Reduce mean time to recovery
• Simplify troubleshooting
• Implement automated remediation of events
• Automate routine processes
• Implement alert filtering