monitoring-capture-service

Guide for using the Grafana MCP to monitor and diagnose the capture service (rust/capture) in production. Use when investigating latency, event loss, Kafka backpressure, Redis issues, rate limiting, Envoy proxy issues, or any capture health question. Covers prod-us and prod-eu environments.

Quality

85%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Quality

Content

82%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The body is a high-quality operational reference: executable MCP calls, exact identifiers, and well-sequenced discovery workflows with a cleanly externalized playbooks file. Its only weaknesses are a few trimmable edge-case details and validation checkpoints that reside in the reference rather than inline.

Suggestions

Move US-only personal-dashboard edge cases (capture-3000-envoy-codes, ws-coarse-lag-explore, 8e93b023-...) and per-VC KMinion instance lists into the reference file to tighten the main overview.

Surface one or two inline validation checkpoints (e.g. a short 'verify after querying' note with the 0.95 acceptance-ratio threshold) so the main body demonstrates the feedback loop without forcing a jump to the reference.

Consider splitting the large Redis-topology and Kafka-topic tables into a second reference file to better distribute the inline reference load.

Dimension	Reasoning	Score
Conciseness	Dense but almost entirely operational reference data Claude does not already know (datasource UIDs, cluster names, topic tables, dashboard UIDs) with no concept over-explanation; not a 5 because a few edge-case details (US-only personal dashboards, per-VC KMinion instances) could be trimmed or externalized.	4 / 5
Actionability	Fully executable guidance: exact MCP tool calls with parameters (e.g. list_prometheus_metric_names with datasourceUid and regex), concrete PromQL scope filters, specific dashboard UIDs, and copy-paste-ready log selectors.	5 / 5
Workflow Clarity	Numbered discovery workflows for each telemetry domain are clearly sequenced, and the referenced playbooks add ordered verification steps with thresholds; not a 5 because explicit validation/feedback loops live in the reference file rather than inline in the main body.	4 / 5
Progressive Disclosure	Well-organized overview with clear section headers and a clearly signaled one-level-deep reference (references/investigation-playbooks.md); not a 5 because only a single external reference exists and a fair amount of inline table content could arguably be split further.	4 / 5
	Total	17 / 20 Passed

Description

87%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

The description is third-person, specific, and explicitly pairs a clear 'what' with a concrete 'Use when...' trigger clause and environment scope. It is a strong, low-conflict description whose only weakness is slightly less than comprehensive keyword/action enumeration.

Dimension	Reasoning	Score
Specificity	Names the domain (Grafana MCP monitoring of the capture service) and concrete actions ('monitor and diagnose') plus specific investigation areas (latency, event loss, Kafka backpressure, Redis issues), with only minor coverage gaps; not a 5 because 'monitor and diagnose' is only two verbs rather than a comprehensive action list.	4 / 5
Completeness	Explicitly answers both what ('Guide for using the Grafana MCP to monitor and diagnose the capture service') and when ('Use when investigating latency, event loss...') with concrete trigger phrases and environment scope.	5 / 5
Trigger Term Quality	Strong natural-language triggers users would actually say ('latency, event loss, Kafka backpressure, Redis issues, rate limiting, Envoy proxy issues, or any capture health question') plus environment names; not a 5 because a few natural variants (e.g. 'ingestion', 'events not being ingested') are absent.	4 / 5
Distinctiveness Conflict Risk	A clear niche — Grafana MCP monitoring of the rust/capture service — with distinct, specific triggers (Kafka backpressure, Envoy proxy, capture health) that minimize overlap with other skills.	5 / 5
	Total	18 / 20 Passed

Validation

100%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: PostHog/posthog
Path: .agents/skills/monitoring-capture-service/SKILL.md
Commit: 6fca5f8

Reviewed: 3 days ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.