flowchart TB Burst["5x ingest burst begins"]:::src G1["Request rate and CPU climb per gateway pod"]:::hpa G2["HPA compares live metric to target"]:::hpa G3["HPA adds gateway replicas"]:::hpa G4["Requests spread across more pods, CPU normalizes"]:::hpa K1["Kafka consumer lag grows on xapi.statements.raw"]:::keda K2["KEDA scaler polls the lag metric"]:::keda K3["KEDA adds processor replicas, each claiming idle partitions"]:::keda K4["Replicas drain backlog over the next minute, lag returns to baseline"]:::keda Burst --> G1 --> G2 --> G3 --> G4 Burst --> K1 --> K2 --> K3 --> K4 classDef src fill:#37474f,stroke:#1c2529,color:#fff,font-size:13px classDef hpa fill:#2a9d8f,stroke:#1f7a6f,color:#fff,font-size:13px classDef keda fill:#7b6cc4,stroke:#544a91,color:#fff,font-size:13px linkStyle default stroke:#999,stroke-width:2px