Repository navigation
feat(muse-code): add Meta Muse Code dashboard template - #417
Merged
Merged
Conversation
Adds a dashboard for Meta's Muse Code terminal coding agent, built from OTel traces emitted by its hook system rather than from its built-in exporter, which cannot authenticate to a third-party backend. 18 panels across five sections: Overview scorecards, Tokens and Cache, Latency, Tool Activity, and Model Behaviour. Spans use gen_ai.* semantic conventions plus muse.* attributes for the agent-specific dimensions (turn outcome, tool status, session end reason). Every panel filters through a multi-select service_name variable, so the board works for teams that split Muse Code telemetry across several service names via MUSE_OTEL_SERVICE.
📊 Dashboard files added in this PR✅ An issue has been created in |
revmag
approved these changes
Sep 15, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Adds a dashboard template for Meta Muse Code, the terminal coding agent Meta shipped in August 2026. This is the CLI agent, distinct from the existing
muse-sparkboard which covers the Model API.Where the telemetry comes from
Muse Code ships a built-in OpenTelemetry exporter, but it cannot be pointed at SigNoz. Its
telemetry.authis a two-value enum,bearerornone, with no field anywhere to supply your own credential, so there is no way to sendsignoz-ingestion-key. Verified on both1.0.3-R2198.1and1.2.1-R2847.1. Separately, the native exporter emits nothing at all in public builds even with a valid login and a parsed config.So this dashboard is built on traces emitted from Muse Code's hook system instead. A hook script receives each lifecycle event as JSON on stdin and posts OTLP/HTTP directly to SigNoz, which works because the script is user code and can set its own headers. The accompanying setup guide is in progress.
Span shape
muse sessionturnturn_id, so hook spans line up with the agent's internaltraceparentchat {model}gen_ai.*attributesexecute_tool {tool}Attributes follow
gen_ai.*semantic conventions, including the dottedgen_ai.usage.cache_read.input_tokensandgen_ai.usage.reasoning.output_tokensspellings already used by themuse-sparkboard, so both dashboards read the same vocabulary. Agent-specific dimensions that have nogen_ai.*equivalent usemuse.*.Panels
Query scoping worth knowing
service.name IN $service_namethrough a multi-select dynamic variable, rather than hardcoding a service. Teams commonly split Muse Code telemetry by repo or team viaMUSE_OTEL_SERVICE.gen_ai.operation.name = 'chat'rather than the span name, because the span name carries the model id and would need a wildcard.gen_ai.tool.name EXISTS.Docs link
The README links to
https://signoz.io/docs/muse-code-monitoring/, which does not resolve yet. The guide is being added in a companion PR to the website repo.