skills.rest
krzko
krzkoCOMMUNITY
Explore all skills by this author →
Website𝕏@krzkoGitHub

metrics-queries

Construct correct Honeycomb queries for OpenTelemetry metrics datasets including gauges, counters, and histograms.

Updated Jun 4, 2026
Search Tags:#metrics#observability#query#opentelemetry#histogram#honeycomb#temporal aggregation
One-click install
npx skills add https://github.com/krzko/pokemon-api --skill metrics-queries-krzko
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: metrics-queries
Source: https://github.com/krzko/pokemon-api/tree/main/.claude/skills/metrics-queries
Command: npx skills add https://github.com/krzko/pokemon-api --skill metrics-queries-krzko

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Querying OpenTelemetry metrics datasets in Honeycomb follows different rules than trace or event datasets, and common operations like bare COUNT, RATE_SUM, or CONCURRENCY fail or return misleading results. This Skill guides you to identify the real metrics dataset, discover per-metric attributes, and apply correct temporal and spatial aggregations. ## Core Features & Use Cases - Dataset and Metric Discovery: Identify true metrics datasets via dataset_type and MetricInfo, then discover metric names and their specific attributes before filtering or grouping. - Correct Aggregation Rules: Learn which operations are forbidden on metrics, the default temporal aggregation per metric type (LAST, INCREASE, SUMMARIZE), and how to override with RATE via calculated fields. - Histogram and Query Math Patterns: Query histogram parent columns or percentile sub-fields, and build formulas for utilization percentages, tail ratios, and error rates. - Use Case: You want per-second request rates from a cumulative counter. The Skill shows you to create a calculated field with RATE($http.server.requests, 300) and apply a spatial aggregation like AVG, avoiding the forbidden RATE_SUM operation. ## Quick Start Ask the AI to query CPU utilization by pod from the Honeycomb metrics dataset and explain which aggregations are valid.

Frequently Asked Questions about metrics-queries

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I query OpenTelemetry metrics in Honeycomb?▼

First identify the metrics dataset via get_environment where dataset_type equals metrics, then use get_dataset_columns to find metric names and attributes. Apply spatial aggregations like AVG, SUM, MAX, or HEATMAP, since bare COUNT and RATE_SUM are forbidden on metrics datasets.

How to calculate per-second rate from a counter metric in Honeycomb?▼

Create a query-scoped calculated field with RATE($metric, range_interval_seconds), then apply a spatial aggregation like AVG or SUM to that field in calculations. RATE_SUM and RATE_AVG are not supported on metrics datasets.

Why does COUNT return meaningless results on metrics datasets?▼

COUNT without a column counts metric events rather than metric values, which is meaningless for metrics. Use AVG, SUM, MAX, or HEATMAP on the metric column instead, or COUNT(metric) with a column specified to count non-null values.

How do I find which attributes are available for a specific metric?▼

Call get_dataset_columns with the metric_name parameter set to your target metric. This returns the resource and data point attributes that co-occur with that metric, since each metric has its own attribute set.

What is the difference between P99 on a histogram parent column and its p99 sub-field?▼

P99(http.server.duration) computes the true 99th percentile across all data points in each step. MAX(http.server.duration.p99) returns the worst pre-computed p99 reported by any single timeseries, useful for worst-case bounds across hosts.

Why do my metrics graphs look spiky or noisy?▼

Spiky graphs usually mean the query granularity is finer than the metric reporting interval, causing uneven data points per bucket. Align granularity with the reporting interval or increase it, and enable Omit Missing Values in the UI for continuous lines.

Related Skills

Explore More Software Engineering Skills→
honeycombiohoneycombio
20

metrics-queries

Query OpenTelemetry metrics datasets in Honeycomb with correct syntax.

Official
Intermediate
Samuelca6399Samuelca6399

observability

Correlate structured logs, metrics, and distributed traces for production monitoring.

Community
Advanced
honeycombiohoneycombio
20

query-patterns

Guide Honeycomb query construction and interpretation for trace and event datasets.

Official
Intermediate
ever-worksever-works
21

next-cache-components

Implement data caching and partial prerendering in Next.js 16.

Official
Advanced