dd-logs

作成者: datadog-labs

ログ管理 - 検索、パイプライン、アーカイブ、およびコスト管理。

npx skills add https://github.com/datadog-labs/pup --skill dd-logs

Datadog Logs

Search, process, and archive logs with cost awareness.

Prerequisites

Datadog Pup (dd-pup/pup) should already be installed:

cargo install --git https://github.com/DataDog/pup

Quick Start

pup auth login

Search Logs

# Basic search
pup logs search --query="status:error" --from="1h"

# With filters
pup logs search --query="service:api status:error" --from="1h" --limit 100

# JSON output is the default
pup logs search --query="@http.status_code:>=500" --from="1h"

Search Syntax

QueryMeaning
errorFull-text search
status:errorTag equals
@http.status_code:500Attribute equals
@http.status_code:>=400Numeric range
service:api AND env:prodBoolean
@message:*timeout*Wildcard

Trace IDs in Log Results

Logs that show a linked trace in the Datadog UI may not include dd.trace_id / dd.span_id in API results. When a trace ID attribute is remapped for trace correlation (via JSON preprocessing or a Trace Remapper processor), the source attribute is removed and the value is stored as an internal attribute that the Logs Search API does not return. The UI's "trace" link reads that internal attribute, so UI and API results differ.

This is expected Datadog Log Management behavior, not a pup bug or an instrumentation problem. Do not retry queries or change instrumentation to "fix" it. Datadog is tracking making these values queryable (support reference FRLOGSS-4306).

Workarounds until then:

  • Emit the trace ID under a separate attribute that is not remapped (e.g. @custom.trace_id) and query that.
  • Pivot the other way: search spans by the log's service/time window via pup traces search, or use the trace link in the Datadog UI.

Pipelines

Process logs before indexing:

# List pipelines
pup obs-pipelines list

# Create pipeline (JSON)
pup obs-pipelines create --file pipeline.json

Common Processors

{
  "name": "API Logs",
  "filter": {"query": "service:api"},
  "processors": [
    {
      "type": "grok-parser",
      "name": "Parse nginx",
      "source": "message",
      "grok": {"match_rules": "%{IPORHOST:client_ip} %{DATA:method} %{DATA:path} %{NUMBER:status}"}
    },
    {
      "type": "status-remapper",
      "name": "Set severity",
      "sources": ["level", "severity"]
    },
    {
      "type": "attribute-remapper",
      "name": "Remap user_id",
      "sources": ["user_id"],
      "target": "usr.id"
    }
  ]
}

⚠️ Exclusion Filters (Cost Control)

Index only what matters:

{
  "name": "Drop debug logs",
  "filter": {"query": "status:debug"},
  "is_enabled": true
}

High-Volume Exclusions

# Find noisiest log sources
pup logs search --query="*" --from="1h" | jq 'group_by(.service) | map({service: .[0].service, count: length}) | sort_by(-.count)[:10]'
ExcludeQuery
Health checks@http.url:"/health" OR @http.url:"/ready"
Debug logsstatus:debug
Static assets@http.url:*.css OR @http.url:*.js
Heartbeats@message:*heartbeat*

Archives

Store logs cheaply for compliance:

# List archives
pup logs archives list

# Archive config (S3 example)
{
  "name": "compliance-archive",
  "query": "*",
  "destination": {
    "type": "s3",
    "bucket": "my-logs-archive",
    "path": "/datadog"
  },
  "rehydration_tags": ["team:platform"]
}

Log-Based Metrics

Inspect log-based metrics:

# List existing log-based metrics
pup logs metrics list

⚠️ Cardinality warning: Group by bounded values only.

Sensitive Data

Scrubbing Rules

{
  "type": "hash-remapper",
  "name": "Hash emails",
  "sources": ["email", "@user.email"]
}

Never Log

# In your app - sanitize before sending
import re

def sanitize_log(message: str) -> str:
    # Remove credit cards
    message = re.sub(r'\b\d{4}[-\s]?\d{4}[-\s]?\d{4}[-\s]?\d{4}\b', '[REDACTED]', message)
    # Remove SSNs
    message = re.sub(r'\b\d{3}-\d{2}-\d{4}\b', '[REDACTED]', message)
    return message

Troubleshooting

ProblemFix
Logs not appearingCheck agent, pipeline filters
High costsAdd exclusion filters
Search slowNarrow time range, use indexes
Missing attributesCheck grok parser
dd.trace_id missing but UI shows a traceExpected: remapped trace IDs become internal attributes (see Trace IDs in Log Results)

References/Documentation

datadog-labsのその他のスキル

dd-audit
datadog-labs
監査証跡の調査 - 誰が何を変更したか、鍵の侵害、コスト急増の根本原因、コンプライアンス証拠(SOC 2/PCI)、およびAIアクティビティの監査。
official
agent-install
datadog-labs
Datadog Operatorを使用してKubernetesにDatadog Agentをインストールします — Single Step Instrumentation (SSI)を有効にする前に必要であり、これは自動的に…
official
agent-observability-auto-experiment
datadog-labs
実在のDatadog LLM-Obsデータに対して、Claude Codeをエージェントとして、反復的なコード改善のヒルクライムをローカルで実行します。ベースライン評価を確立し、1つの…
official
agent-observability-eval-bootstrap
datadog-labs
本番トレースから評価器を初期設定する — デフォルトではオンラインLLM-judge評価器を提案し、確認後にDatadogで無効なドラフトとして作成します…
official
agent-observability-eval-pipeline
datadog-labs
計装されたml_app向けのエンドツーエンドのエージェント可観測性パイプライン — 本番トレースを分類し、障害の根本原因を特定し、評価器をブートストラップし、その後(オプションで)…
official
agent-observability-experiment-analyzer
datadog-labs
LLM実験結果を分析します。単一または比較実験、探索的またはQ&Aモードに対応します。ユーザーが「analyze experiment」「compare…」と言った場合に使用します。
official
agent-observability-replay-trace
datadog-labs
開発者が、出力が気に入らなかった特定のAgent Observability / LLM Obsトレースに対して反復作業を行いたい場合に使用します — そのトレースを再実行して…
official
agent-observability-trace-rca
datadog-labs
本番環境のLLMトレースに対する根本原因分析。LLMアプリケーションが失敗する理由を診断し、評価ジャッジの判定、ランタイムエラー、または構造的な問題から分析を開始します…
official