Skip to the rows

agentsclimarketplace

Observability engineering

Skill valdomirosouza/agent-skills/skills/observability-engineering

Expert AI Agent Skills for SDD, SRE, DevSecOps and Enterprise Engineering

Install

npx -y skills add valdomirosouza/agent-skills --skill observability-engineering

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Implements production observability using the three pillars (logs, metrics, traces), defines and monitors SLI/SLO/SLA, and instruments services with OpenTelemetry. Use when adding observability to a service, defining SLOs, configuring alerts, designing dashboards, or setting up telemetry pipelines. Also use when asked about Golden Signals, error budgets, structured logging, trace context propagation, or OTel configuration.

SKILL.md

4.2 KB, as published. Nobody here has run it

Observability Engineering

Contents

Golden Signals (mandatory for every service)
Structured logging standard
Metrics conventions
Distributed tracing (OpenTelemetry)
SLI / SLO / SLA definitions and templates → sli-slo-templates.md
Telemetry stack reference
Instrumentation code patterns → instrumentation-guide.md

Golden Signals — Mandatory for Every Service

Signal	Definition	Example metric	Base alert
Errors	Rate of failed requests (4xx/5xx)	`http_requests_total{status=~"5.."}`	> 1% for 5 min
Latency	Response time — always track p50/p95/p99	`http_request_duration_seconds`	p99 > SLO threshold
Traffic	Request/operation volume per second	`http_requests_total`	Anomaly by std deviation
Saturation	Resource utilization (CPU, mem, disk, queue)	`container_cpu_usage_seconds_total`	> 80% for 10 min

Structured Log Standard

Every log entry MUST contain:

{
  "timestamp": "2024-01-15T10:30:00.123Z",
  "level": "ERROR",
  "service": "payment-service",
  "version": "1.4.2",
  "environment": "production",
  "trace_id": "4bf92f3577b34da6a3ce929d0e0e4736",
  "span_id": "00f067aa0ba902b7",
  "message": "Payment processing failed",
  "http": { "method": "POST", "path": "/v1/payments", "status_code": 504, "duration_ms": 3001 }
}

Required fields: timestamp (ISO 8601 UTC), level, service, version, trace_id, span_id, environment

Forbidden in logs: passwords, tokens, API keys, CPF, card numbers, raw request bodies with PII

Metrics Naming Convention

[namespace]_[subsystem]_[metric]_[unit]

# Examples:
http_server_requests_total{method, path, status}       # Counter
http_server_request_duration_seconds{method, path}     # Histogram
db_connections_active{pool, database}                  # Gauge
queue_messages_pending{queue_name, consumer_group}     # Gauge

Label rules:

High-cardinality labels are forbidden (user_id, IP, UUID)
Labels must be stable and pre-defined in spec
Max 10 labels per metric

Distributed Tracing — OpenTelemetry Standard

standard: OpenTelemetry (OTel)
propagation: W3C TraceContext + Baggage

required_spans:
  - All incoming HTTP requests
  - All outgoing calls (external services, DB, queues)
  - Critical business logic operations

required_attributes:
  - service.name, service.version, deployment.environment
  - http.method, http.url, http.status_code
  - db.system, db.statement (sanitized — no PII)
  - error (boolean + error.message when applicable)

sampling:
  production: tail-based — 10% normal, 100% errors
  staging: 100%

For code examples → instrumentation-guide.md

SLI / SLO / SLA

Term	Definition	Owner
SLI	Quantitative measure of service behavior	Engineering
SLO	Internal reliability target	Engineering + Product
SLA	Contractual commitment to customer	Business + Legal
Error Budget	`1 - SLO target` — allowed margin of failure	Engineering

Key rule: Internal SLO must be stricter than external SLA (buffer of safety). Example: SLA = 99.5% → SLO = 99.9%

For slo.yaml template and alert configuration → sli-slo-templates.md

Telemetry Stack Reference

Layer	Options
Collection	OpenTelemetry Collector, Fluent Bit
Logs storage	Loki, Elasticsearch, OpenSearch
Metrics storage	Prometheus + Thanos (long-term)
Traces storage	Tempo, Jaeger, Zipkin
Visualization	Grafana, Kibana
Alerting	AlertManager → PagerDuty / OpsGenie
AIOps	Grafana ML, Dynatrace Davis (anomaly detection)

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.