Enable Data Streams Monitoring

Data Streams Monitoring runs inside the same Datadog SDK that APM already injected. It adds a small context header to each message so Datadog can stitch producers, queues, and consumers into pipelines and measure end-to-end latency and lag. There is no new agent, no new package, and no code change for supported libraries. It is one tracer setting per service.

Before doing anything else: Fully resolve all variables in ## Context to resolve before acting, then get the user's explicit yes in Step 1. Do not change any configuration before that yes.

Triggers

Invoke this skill when:

• The user asks to enable Data Streams Monitoring, monitor queue or consumer lag, see pipeline latency, or monitor message flow through Kafka / RabbitMQ / SQS / SNS / Kinesis / Pub/Sub
• enable-ssi (Kubernetes or Linux) reached its event-driven check and found a fit
• The user's system is event-driven: microservices that hand work to each other asynchronously, services that turn a synchronous request into async background work, or Lambda functions triggered by queues or streams

Do NOT invoke this skill if:

• APM is not yet set up on the target services, unless enable-ssi is calling this skill after applying its SSI config. Otherwise run the dd-apm install flow first; DSM needs the Datadog SDK running in the process
• The user only wants Kafka broker or cluster health (brokers, topics, partitions, configs). That is Kafka Console, an Agent check, and out of scope for this skill
• The user already declined DSM in this session. Do not ask again