OpenTelemetry Collector Pipeline Designer
Designs OpenTelemetry Collector pipeline configurations with receivers (otlp, prometheus, filelog), processors (batch, attributes, tail_sampling), and exporters (otlphttp, jaeger, elasticsearch).
npx skills add agentskillexchange/skills --skill otel-collector-pipeline-designer
The OpenTelemetry Collector Pipeline Designer skill generates comprehensive otel-collector configuration files for telemetry data routing and transformation. It configures receivers including otlp (gRPC/HTTP), prometheus for metric scraping with scrape_configs, filelog for structured log ingestion with multiline parsing, and hostmetrics for system-level CPU/memory/disk collection.
What this skill actually does
The skill designs processor chains with batch processor for throughput optimization with send_batch_size and timeout tuning, attributes processor for span/metric attribute manipulation (insert, update, delete, hash operations), memory_limiter for backpressure management, and tail_sampling for intelligent trace sampling based on latency, error status, and string attribute policies.
Advanced pipeline configurations include multi-pipeline fanout for sending traces to both jaeger and otlphttp exporters simultaneously, connector usage for traces-to-metrics (spanmetrics connector) and logs-to-metrics (count connector) derivation, and extension configurations for health_check, pprof, and zpages debugging endpoints. The skill generates Kubernetes-native configurations using the OpenTelemetry Operator CRDs (OpenTelemetryCollector kind) with DaemonSet and Deployment mode selection, and produces Helm values.yaml for the opentelemetry-collector chart deployment.