AIOPS in DevOps

Explore the foundations of IT operations, including monitoring, incident management, and security, and learn how AI can improve service availability, performance, and cost efficiency.

Explore IT operations challenges at scale, complexity, and limited resources, from diverse systems and global teams to keeping high availability with just-in-time provisioning, and see how AI-powered automation supports AIOps.

Explore how artificial intelligence integrates with IT operations to automate routine tasks and forecast capacity, outages, and root causes from telemetry data.

Explore it operations use cases in ai ops, including monitoring, alerting, security and compliance, capacity planning, and helpdesk automation, with incident root cause analysis, self-service desk, and service load forecasting.

Set up the AIOPS in DevOps development environment and exercise files using Python 3.7, Visual Studio Code, and a virtual environment, then install and run Jupyter notebooks.

Explore how deep learning elevates classification in IT operations, enabling root cause predictions of system incidents for faster, proactive resolutions using supervised models.

Root cause analysis identifies fundamental reasons behind IT incidents by tracing symptoms to their origin, with AI models predicting likely root causes to accelerate diagnosis.

Learn how a binary symptom dataset from IT ops incidents, with indicators like high cpu load and network latency, informs a classical RCA classification model to predict root causes.

Load and pre-process the RCA dataset for machine learning by converting the root cause to numeric, converting to numpy arrays, and applying one-hot encoding, preparing for a Keras model.

Build a softmax model with three dense layers in Keras to predict root causes from symptoms, 20 epochs, batch 100, 20% validation, Adam optimizer, and categorical cross-entropy, achieving 81% accuracy.

Predict root causes of incidents using a Keras-based RCA model, feed symptom features, and perform batch predictions to identify database issues, with hyperparameter tuning for speed and accuracy.

Discover how generative AI creates content from large data sets and enhances AIOps, including ticket triage, root cause analysis, and personalized user interactions, with governance and privacy challenges.

Discover how large language models drive generative ai in ai ops and devops. Use prompts to summarize incidents, classify tickets, power chatbots, and retrieve knowledge from logs and docs.

Explore generative ai use cases in ai ops and show how logs and a format enable a prompt to a language model that extracts timestamps, error codes, and ip addresses.

Leveraging generative ai, synthesize incident data from logs, tickets, and chat transcripts into cohesive summaries that outline chronology, impact, root causes, and resolution, accelerating reports for leadership and stakeholders.

Explore a self-help chatbot use case for ai ops that delivers context-aware answers by querying a vector-based knowledge base built from vendor manuals and internal docs.

Generative AI speeds up scripting for IT operations by turning natural language prompts into working scripts, with built-in tests and validation before production deployment.

This lecture demonstrates code generation with the OpenAI API to create a Python function that summarizes an Apache access log by response code using ChatGPT 3.5 Turbo.