Mike Vizard details how OpenTelemetry is expanding to support observability and automation for AI agents, showing how Cisco and Splunk are integrating these capabilities to streamline IT operations.

OpenTelemetry Extensions Enhance Observability for AI Agents

OpenTelemetry is expanding its capabilities by adding support for observability of AI agents. The maintainers of this open source telemetry project are developing semantic extensions that enable platforms to collect and analyze data from AI agents alongside traditional application telemetry.

According to Morgan McLean (OpenTelemetry co-founder and Cisco/Splunk product director), these extensions will unify observability efforts and reduce the need to deploy separate monitoring platforms for AI agents. As OpenTelemetry has become easier to deploy and configure—such as through automatic data source discovery—it has seen growing adoption in DevOps environments.

Recent Developments:

  • Cisco previewed AI agents in its Splunk Observability suite. These agents leverage OpenTelemetry to automate the collection, issue detection, root cause analysis, and resolution tasks within IT environments.
  • Cisco launched a Data Fabric platform to aggregate and analyze machine data at scale, which soon will include time-series analysis powered by various AI toolkits.
  • The AI Canvas initiative from Cisco involves training large language models (LLMs) with telemetry data to automate and unify network, cloud, security, and observability workflows.

Why This Matters: Maintaining observability in complex, rapidly evolving IT landscapes is increasingly difficult using manual tools. With OpenTelemetry’s enhancements and pervasive use of AI agents, IT teams can gain unified visibility, automate time-consuming tasks, and focus on decision validation rather than manual intervention.

Key Points:

  • OpenTelemetry’s semantic extensions support AI agent telemetry
  • Cisco and Splunk drive innovation in unified observability and IT automation
  • New platforms enable aggregation, analysis, and automated response for a complex, data-rich IT environment
  • AI-driven solutions help replace legacy monitoring with proactive operational intelligence

For more detail, see the original announcement by Mike Vizard and Cisco’s AI Canvas announcement.

This post appeared first on “DevOps Blog”. Read the entire article here