In addition to Agent Timeline, Canvas Agent and Skills supercharge engineering teams and their AI agents with a shared production observability layer

Honeycomb.io, the observability platform built for the new shape of software, introduced a series of agentic intelligence and agent observability features purpose-built for AI agents in production. These new and enhanced capabilities include Agent Timeline, Canvas Agent, and Canvas Skills. Engineering teams will now have real-time visibility into what their agents are actually doing without proprietary software development kits (SDKs) or framework lock-in.

AI agents are rapidly handling tasks such as code generation, incident triage, cloud infrastructure deployment, and customer service. However, the tools built to observe software systems weren’t designed for non-deterministic, multi-hop agent workflows. Dashboards break down, averages lie, and when an agent causes an incident, teams have no way to reconstruct what it decided or why.

“AI has upended how software works: introducing real nondeterminism into production, and fundamentally changing how teams of humans and agents build and validate code,” said Christine Yen, cofounder and CEO of Honeycomb. “Engineers are drowning in uncertainty as most observability tools weren’t built for this sort of “unknown unknown.” Honeycomb was built for the hardest parts of building software—and autonomous agents have taken those hardest parts mainstream.”

Marketing Technology News: MarTech Interview with Kurt Donnell, CEO @ Freestar

Observability for AI systems

Honeycomb’s new AI-native observability capabilities allow Engineering teams to track and analyze the performance, behavior, and interactions of AI agents. With Agent Timeline, users can render multi-agent, multi-trace workflows as a single coherent view, connecting every LLM call, tool invocation, agent handoff, and downstream system impact in real time. Engineering teams can now trace what an agent did, reconstruct the full decision path, and understand failures without switching tools or piecing together logs by hand.

Observability with AI agents
New and enhanced features from Honeycomb supercharge human and agent teams by surfacing intelligence at machine speed. These include:

• Canvas, newly rebuilt to serve as a collaborative workspace, a chat interface, and autonomous agent all in one, allows engineering teams to query issues in plain English, work with human and agent team members on investigations, and produce sharable visualization snapshots.
• Auto-investigations set your Canvas agent to work automatically when an alert fires, an SLO burns, or an anomaly surfaces, gathering data, creating and testing hypotheses, and proposing remediation, all before an engineer even opens their laptop.
• Canvas Skills encode your best engineers’ debugging knowledge and best practices for frameworks or services like Kafka into reusable playbooks that run autonomously.

Marketing Technology News: Programmatic Ad Platforms With Unique AdTech Features

“Honeycomb is great at dealing with unknown-unknowns, and Canvas takes it even further,” said Shogo Wada, Staff Software Engineer at Bubble. “During a recent investigation, we found a cause of slowness in our API by asking Canvas to investigate. While the cause was not directly visible on the affected spans, Canvas compared whole traces and found patterns within their child spans. Before Canvas, this would have been a manual process of opening traces one by one. Additionally, as Canvas is a collaborative tool among the team, we could theoretically explore multiple hypotheses in parallel, getting to a root cause faster.”

Built on open standards for maximum flexibility
As a prominent supporter and contributor to the OpenTelemetry project, Honeycomb has integrated the OpenTelemetry (OTel) GenAI semantic conventions (v1.40.0) into its platform. By making gen_ai.* attributes first-class citizens, Honeycomb ensures that model evaluations, tool executions, MCP calls, LLMs, and agents are all properly observed. This alignment with OpenTelemetry GenAI standards allows for the automatic generation of structured GenAI insights without requiring re-instrumentation as specifications change, custom setups, or the use of proprietary SDKs.

Honeycomb’s proprietary unified data store was built for exactly the kind of unpredictable, high-dimensionality telemetry that AI agents generate, from the ground up.