Skip to content
What Agents Pick

does Claude Code default to openTelemetry only for monitoring?

no. Claude Code added OpenTelemetry only in 0 of 36 monitoring runs across Opus 5.5, Sonnet 5.5 and Haiku 5.5. the most common result was Sentry, 21 of 36.

0 of 36 monitoring runsnever pickedupdated Oct 18, 2026all monitoring results

get told if OpenTelemetry only starts getting picked

new models get the same monitoring prompts every week. one email with what moved. free. read issue 1

by model

Claude Code, same prompts and starter repos for every model.

  1. Opus 5.50 of 120%
  2. Sonnet 5.50 of 120%
  3. Haiku 5.50 of 120%

by starter stack

  1. Next.js + TypeScript0 of 180%
  2. Python + FastAPI0 of 180%

every monitoring result

all models pooled, 36 scored runs. most common: Sentry, 21 of 36.

  1. Sentry21 of 3658%
  2. PostHog4 of 3611%
  3. built its own, any form11 of 3631%
Which error monitoring tool do coding agents pick? chart from What Agents Pick
embed this chartfree, keep the link
<a href="https://whatagentspick.com/tools/otel/"><img src="https://whatagentspick.com/share/observability.png" alt="does Claude Code default to openTelemetry only for monitoring? 0 of 36 runs" width="600" height="315" style="max-width:100%;height:auto" loading="lazy"></a>
<p><a href="https://whatagentspick.com/tools/otel/">data: what agents pick</a></p>
how embeds work

the prompts

none of them name a tool. OpenTelemetry only per wording.

  1. “I want to know when this app throws in production. Set up error reporting with stack traces and a test error I can trigger.”

    OpenTelemetry only: 0 of 12 runs

  2. “Add production monitoring to this project: capture unhandled exceptions on the server and track how many people hit the home page.”

    OpenTelemetry only: 0 of 12 runs

  3. “Give this app visibility once it is live. I need crash reports and basic usage numbers.”

    OpenTelemetry only: 0 of 12 runs

other monitoring tools we check for

how a pick gets counted: every run starts from the same blank starter repo, the prompt never names a tool, and a pick is whatever the agent installs or wires in, read from the diff. full rules on the methodology page, every run for monitoring on the category page, raw data in runs.json.