Skip to content
Skills Library
DevOps & cloud · Analysis

distributed-tracing

/distributed-tracing

Follow one request across service boundaries using spans, propagated context, and critical-path reading to locate where latency and errors originate. Use when a request is slow or failing and the cause lives between services, not inside any single one.

Repository
AI-SKILLS ↗
Licence
MIT
Last repository change
05.09.2026

Assessment and security

90/ 100Class: good
Format25 / 25
Frontmatter with name and description, valid name, description up to 1,024 characters, name matching the folder.
Discoverability25 / 25
Description explains when to use the skill, has a reasonable length and uses third person.
Content25 / 35
At least 150 words and at most 500 lines, headings, step lists, examples or code.
Hygiene15 / 15
No placeholders such as TODO, no Windows paths, with a licence.
Riskno signals detected in the scan
Files in this package
1
Scripts
0
HTML comments
0
URLs in SKILL.md
0
Deprecation

No signals detected. Repository changed on 05.09.2026, without references to retired models or APIs.

Licence

MIT, open: use and modify while retaining author attribution.

What it does and when to use it

Follow one request across service boundaries using spans, propagated context, and critical-path reading to locate where latency and errors originate. Use when a request is slow or failing and the cause lives between services, not inside any single one.

Original author description

Follow one request across service boundaries using spans, propagated context, and critical-path reading to locate where latency and errors originate. Use when a request is slow or failing and the cause lives between services, not inside any single one.

How it works

  1. Method
  2. Signals
  3. Boundaries
Original SKILL.md headings · 48 lines, 407 words
SKILL.md preview
---
name: distributed-tracing
description: Follow one request across service boundaries using spans, propagated context, and critical-path reading to locate where latency and errors originate. Use when a request is slow or failing and the cause lives between services, not inside any single one.
---

# Distributed tracing

In a mesh of a dozen services, "checkout is slow" names a symptom with no
address. A log gives you one service's account; a trace gives you the whole
conversation with timing attached. Profile each service alone and you polish
the fast ones while the time drains into a call three hops down you never saw.

## Method

1. **Propagate context or the trace snaps at hop one.** Verify every service
   forwards the `traceparent` header (W3C Trace Context) or B3. A dropped
   header spawns a fresh root, and the child spans orphan into a tree you can
   never rejoin to the parent.
2. **Name spans for the operation, not the method.** `handleRequest` says
   nothing; `GET /orders/:id` and `SELECT orders by user` say what ran. Set
   `span.kind` to server, client, or producer so the waterfall nests parent
   and child correctly.
3. **Follow the critical path, not the sum.** In the waterfall, trace the chain
   of spans a later span actually blocked on. A 400ms span running beside a
   900ms sibling is free; the 900ms one owns the budget. Attack the longest bar
   on the blocking path.
4. **Ride baggage for cross-cutting facts.** Put `user.tier` or `feature.flag`
   into OpenTelemetry baggage so it propagates to every downstream span. Later
   you query "traces where tier=free and duration>2s" without threading that
   value through every signature by hand.
5. **Hunt the gap between spans, not only the long ones.** Dead time with no
   child span open is queueing, pool wait, or GC that your instrumentation
   missed. A 300ms hole before the database span is the connection pool
   starving, not the query executing.
6. **Sort errored traces by where status flipped.** Filter `status=error` and
   find the deepest span that set it. Parents inherit the failure, so the leaf
   that first went red is the origin; read its events for the exception.

## Signals

- Can you point to the one span on the critical path that owns most of the
  latency?
- Does a single trace id ride the request from edge to database?
- On a failed request, did you locate the deepest span that first set error?

## Boundaries

Tracing shows where time and errors move between services; it will not explain
why one function is slow inside its own span. For that, drop to a profiler or
add metrics. Sampling means a given one-off trace may not exist at all:
tail-based sampling keeps errors, head-based may discard them.

Versions

RepositoryScoreLicence
This page · Amey-Thakur/AI-SKILLS90MIT
rootcastleco/rei-skills100MIT
wshobson/agents100MIT
rmyndharis/antigravity-skills100MIT
HermeticOrmus/LibreUIUX-Claude-Code100MIT
agent-skills-hub/agent-skills-hub100MIT
arjunprabhulal/devops-skills90MIT

Database snapshot: 13.11.2026 · package version a2a1712a50bc (SHA-256 digest of the entire package)

Downloads in the last 30 days: 0 · ranking

Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.

Similar skills

  • 100setting-up-distributed-tracing
    Execute this skill automates the setup of distributed tracing for microservices. it helps developers implement end-to-end request visibility by configuring context propagation, span creation, trace collection, and analysis. use this skill when the user re... Use when appropriate…
  • 100setting-up-distributed-tracing
    Execute this skill automates the setup of distributed tracing for microservices. it helps developers implement end-to-end request visibility by configuring context propagation, span creation, trace collection, and analysis. use this skill when the user re... Use when appropriate…
  • 90mesh-observe
    Design service mesh observability — distributed tracing, service-level metrics, and dashboards. Use when asked to "add service mesh observability", "set up distributed tracing in the mesh", or "build service-level dashboards".
  • 100matlab-instrument-opentelemetry-tracing
    Add OpenTelemetry tracing to MATLAB code. Use when the user asks to "add tracing", "instrument with spans", "add OpenTelemetry", "trace my code", "add observability" (when about tracing), or mentions "spans", "distributed tracing", or "OTel tracing" in the context of MATLAB…
  • 90distributed-debugging-debug-trace
    Set up debugging environments, distributed tracing, and diagnostic tools, and establish troubleshooting practices for development and production. Use when configuring debugging workflows or tracing requests across services.
  • 100dt-obs-tracing
    Distributed traces, spans, service dependencies, and request flow analysis. Use when investigating span-level details, failures, performance bottlenecks, or trace correlation. Trigger: "trace analysis", "slow requests", "failed spans", "service dependencies", "distributed…

Feedback and reports

Your message goes to Bartosz Sokoliński. An email address is needed only for a reply. See the privacy policy.