• Gabriel Reid Package delivery services like UPS, FedEx, and the postal service have a tough job. • They contend with a never-ending stream of packages to be delivered, with expectations of prompt and reliable delivery. • In building the Datadog Log Forwarding feature, we had to contend with similar challenges. • Running a package delivery service initially sounds relatively simple, but consider details like: - How do you use resources efficiently to get packages delivered? • - How do you avoid wasting time revisiting an address when no one is there to receive the package? • - What do you do if a package has a bad or non-existent address?
Article Summaries:
- Datadog has released a new Log Forwarding feature that treats log delivery like a large‑scale package service. The system ingests logs as schemaless JSON, transports them on Kafka partitions, and forwards them to any external HTTP endpoint-including ElasticSearch, Splunk, or custom services. To handle thousands of unpredictable destinations, Datadog groups logs per destination across multiple partitions, then regroups them for delivery. The design prioritizes low latency, high throughput, and reliability, even when endpoints are slow, flaky, or unavailable. The post explains how the team applied package‑delivery analogies to solve challenges of efficient routing, avoiding wasted retries, and ensuring successful log delivery.
Sources: