AWS re:Invent 2025 - AI-Native Era of Observability: How You Can Get Started Today (AIM220)

The AI-Native Era of Observability: How to Get Started Today

The Evolving Software Landscape

• Software complexity has increased dramatically in recent years
• Monolithic architectures have given way to microservices, multi-cloud, and serverless environments
• This has made it much more difficult to understand and manage software systems

The Observability Challenge

• Telemetry data (logs, metrics, traces) has exploded, growing 100x in just 8 years
• More data does not necessarily mean better observability or reliability
• Relying on dashboards, alerts, and manual investigation has become overwhelming
• 92% of dashboards are only used for 1 week, yet more are continually added

The Limitations of Traditional Observability

• Observability is a passive term - the goal should be reliability, stability, and business insights
• Companies have invested heavily in observability tools, but have not seen commensurate improvements in reliability
• The data growth outpaces the ability to effectively manage and derive value from it

Introducing Oolie: The Autonomous Observability Agent

• Oolie is the world's first autonomous observability agent, powered by advanced AI and machine learning
• Oolie leverages Coralogix's massive observability data platform to:
  • Automatically investigate issues and incidents
  • Correlate disparate data sources to identify root causes
  • Eliminate noise and focus on the most relevant signals
  • Provide actionable insights and recommended fixes

How Oolie Works

• Oolie is not a single AI model, but a team of specialized agents powered by small language models (SLMs)
• These agents collaborate to conduct investigations, analyze data, and generate insights
• Oolie learns the specifics of the customer's observability data and environment to provide context-aware analysis
• The key components are the model, system prompt, agent architecture, knowledge base, and evaluation/evolution

Real-World Results

• Example: Oolie helped a customer with a notification service experiencing random latency spikes
  • Oolie identified connections to other services and the underlying RDS database
  • Oolie found that unindexed tables in the RDS database were causing the issues
  • The fix was straightforward once the root cause was identified

The Future of Observability

• Oolie represents a paradigm shift, moving from linear observability improvements to exponential gains in reliability
• The next steps are proactive alerting and remediation, followed by the "holy grail" of preventive observability
• Oolie is now generally available to all Coralogix customers, with the proactive capabilities coming in the next few months

Key Takeaways

• Software complexity has outpaced traditional observability approaches
• Autonomous, AI-powered observability can provide 10x improvements in reliability
• Oolie is a pioneering solution that leverages specialized AI agents to investigate, analyze, and resolve issues
• Adopting AI-native observability is crucial to keep pace with the evolving software landscape