feat: 🌈 friday vibe coding

Author: fungiboletus

.claude/settings.json

@@ -22,15 +22,15 @@ This file provides guidance to Claude Code (claude.ai/code) when working with co
 
 ## Architecture Overview
 
-SensApp is a **sensor data platform** built with Rust that scales from edge deployments (SQLite) to big data clusters (ClickHouse). It uses an **event-driven architecture** with an internal message bus for component communication.
+SensApp is a **sensor data platform** built with Rust that scales from edge deployments (SQLite) to big data clusters (ClickHouse). It uses **direct storage calls** for simplicity and horizontal scalability.
 
 ### Core Components
 
 #### Data Ingestion (`src/ingestors/`)
 
 - **HTTP REST API** with Axum web framework
-- **MQTT client** for IoT device integration
 - **Multi-format support**: JSON, CSV, SenML, InfluxDB line protocol, Prometheus remote write
+- For message queue protocols (MQTT, Kafka, etc.), use external tools like Telegraf or Vector to bridge to HTTP
 
 #### Storage Abstraction (`src/storage/`)
 
@@ -60,13 +60,12 @@ SensApp is a **sensor data platform** built with Rust that scales from edge depl
 
 - Primary config file: `settings.toml`
 - Database connections configured via connection strings
-- MQTT configurations support multiple endpoints
 - Sentry integration for error tracking and monitoring
 
 ### Key Design Patterns
 
 - **Async-first** using Tokio runtime throughout
-- **Event-driven** architecture with message passing
+- **Stateless architecture** with direct storage calls for horizontal scalability
 - **Storage-agnostic** design with trait-based abstractions
 - **Type safety** for sensor data with compile-time guarantees
 - **Scalable deployment** patterns from single-node to distributed clusters, but the complexity is handled by the storage layer (the databases).

CLAUDE.md

@@ -91,11 +91,7 @@ clap = { version = "4.5", features = ["derive"] }
 serde-inline-default = "0.2"
 serde_bytes = "0.11"
 sentry = { version = "0.42", features = ["anyhow", "tower"] }
-rumqttc = { version = "0.24", features = ["url", "websocket"] }
 url = "2.5"
-rand = "0.9"
-#geobuf = "0.1.4"
-#protobuf = "3.0.2"
 futures-lite = "2.6"
 #gcp-bigquery-client = { git = "https://github.com/lquerel/gcp-bigquery-client.git", rev = "107a5557df6336933f6b0bcf330aa91fe6ca866a" }
 gcp-bigquery-client = { version = "0.26.0", optional = true }

Cargo.lock

@@ -7,7 +7,7 @@ It enables the handling of small time series data of the edge efficiently to lar
 ## Features
 
 - **Flexible Time Series DataBase Storage**: Supports various time-series databases such as SQLite, PostgreSQL (with optional TimeScaleDB plugin), DuckDB, ClickHouse, BigQuery, and RRDCached, with the potential to extend support to other databases in the future.
-- **Multiple Data Ingestion Protocols**: Easy data ingestion via HTTP REST API and MQTT.
+- **HTTP REST API**: Easy data ingestion via HTTP REST API with support for multiple data formats.
 - **Compatibility with Existing Pipelines**: Offers Prometheus Remote Write and InfluxDB line format support for seamless integration into existing sensor data pipelines.
 - **Data formats**: Supports various data formats like JSON, CSV, SenML, and Apache Arrow IPC.