Observability with OpenTelemetry & Elastic

Still relying on `print()` for debugging? Learn how OpenTelemetry automatically instruments your Python apps for powerful, industry-standard observability.

#1about 1 minute

The growing need for observability in complex applications

Modern applications with many moving parts, like those in the GenAI space, require robust monitoring to diagnose and fix issues effectively.

#2about 4 minutes

Moving beyond print statements for Python monitoring

While print() statements are a common starting point for debugging, Python's native logging module offers a more structured, albeit limited, approach.

#3about 5 minutes

Introducing OpenTelemetry as a universal standard

OpenTelemetry provides a vendor-agnostic, open-source framework for instrumenting applications to emit telemetry data for analysis.

#4about 7 minutes

Exploring the three main signals: traces, metrics, and logs

Observability is built on three core signal types: traces for request paths, metrics for numerical data, and logs for event records.

#5about 5 minutes

Using manual and automatic instrumentation in your code

You can add OpenTelemetry to your application by manually inserting code snippets or by using automatic instrumentation for common libraries and frameworks.

#6about 3 minutes

Combining OpenTelemetry data with the Elastic stack

Elastic natively supports the OpenTelemetry protocol and schema, allowing you to collect, store, and visualize telemetry data in a centralized platform.

#7about 3 minutes

Visualizing application performance with an Elastic dashboard

A live demonstration shows how an OpenTelemetry-instrumented application sends data to Elastic, revealing metrics like latency, throughput, errors, and logs.

#8about 2 minutes

Why observability is critical for Python and AI applications

Adopting observability standards like OpenTelemetry is crucial for Python developers to monitor, debug, and optimize increasingly complex AI and LLM-based systems.