adding baggage to whitepaper

Author: gyliu513

whitepaper.md

@@ -23,6 +23,7 @@ See [the contributing section](#contributing) on proposing changes to this paper
   * [Logs](#logs)
   * [Traces](#traces)
   * [Profiles](#profiles)
+  * [Baggage](#baggage)
   * [Dumps](#dumps)
 * [Correlating Observability Signals](#correlating-observability-signals)
   * [Signal correlation](#signal-correlation)
@@ -259,6 +260,17 @@ Profiling data produced by runtimes typically includes statistics down to the li
 
 > Image 2 shows an example icicle graph for a CPU profile of an application written in Go. It includes a kernel operation done by the Syscall call. This profile highlights that 35% of CPU time was used to compute hash, indicating potential optimization opportunities. [Source](https://pprof.me/9ce2c2d/).
 
+### Baggage
+
+OpenTelemetry Baggage is a mechanism for propagating small key-value pairs throughout a distributed system. It allows you to attach contextual information to a request as it flows through various services. While Baggage is not primarily designed for debugging or monitoring like the other signals, it complements them by enabling advanced context-sharing capabilities.
+
+Baggage consists of:
+- Key-Value Pairs: Lightweight metadata (e.g., user_id=12345, region=us-west-1).
+- Global Context: Shared across services as part of the request context.
+- Cross-Service Propagation: Automatically carried by OpenTelemetry propagators.
+
+Checkout [OpenTelemetry Baggage](https://opentelemetry.io/docs/concepts/signals/baggage/) for detail.
+
 ### Dumps
 
 In software development, core dump files are used to troubleshoot a program, i.e., a crashed process. Classically, the operating system, with the help of some configuration such as location, name convention or file size, writes an image of the process's memory at the time of the crash for analysis. In cloud-native, however, core dump files' collection of a large cluster can easily create a bottleneck in storage or even the network, depending on how the cluster's storage is attached to the cluster nodes. For example, processing-intensive applications could generate core dump files of double-digit Gigabyte size.