====== Stream Processing (Kinesis Data Analytics) ======

**What it is:** Continuous processing of events as they arrive in a stream.

**What it’s for:**
  * Real-time calculations (aggregations, filtering, anomaly detection).
  * Produce up-to-date outputs for dashboards or APIs.

**Key ideas:**
  * Streaming is continuous (unbounded data).
  * Results can be written to a datastore for API reads (design-specific).

**Exam cues:**
  * “real-time analytics on events” → stream processing (KDA).

**Hard words:**
  * *unbounded* /ʌnˈbaʊndɪd/: không có điểm kết thúc (dòng dữ liệu liên tục)
  * *aggregation* /ˌæɡrɪˈɡeɪʃn/: tổng hợp (sum/count/avg)
  * *anomaly* /əˈnɑːməli/: bất thường
  * *filtering* /ˈfɪltərɪŋ/: lọc