Skip to content
On this page


wallowa wraps DuckDB with a CLI, web UI, and fetcher to make it quick and easy to gain insight into your SDLC. As much data handling as possible is delegated to DuckDB. The tool is written primarily in Rust and deploys as a single statically linked binary. Check out the many thanks to the projects that this project depends on.

There are several components.

Architecture diagram

View the MermaidJS source code for the diagram.
    subgraph wallowa binary
      subgraph cli
      subgraph new
      subgraph db
      subgraph fetch
      subgraph web
      subgraph sources
        GitHub[GitHub REST API]
    cli --> fetch
    cli --> web
    cli --> new
    fetch --> sources
    web --> sources
    sources --> db


The core of the tool is DuckDB. Database utilities and migration/DDL statements can be found in src/


Data is queried directly from wallowa_raw_data using DuckDB json extraction functions. See the src/github/ function for an example of what this looks like in practice.

There is a design tradeoff to querying the raw data directly. The queries themselves are more complicated and expensive to run compared to the data being loaded into normalized tables. However, this leaves a great deal of flexibility to adapt the queries to specific questions that require more of the API payload data. Both styles can be mixed, but for now the reduced complexity approach is being used. You can create views from the more complex queries as a middle ground, too.

Entity relationship diagram

Entity relationship diagram

View the MermaidJS source code for the diagram.

The MermaidJS source for the diagram.

        int id
        timestamp created_at
        timestamp loaded_at
        string data_source
        string data_type
        json metadata
        string data
        string name
        string value


Sources represent the data sources used to drive insights through queries and charts.

Sources are responsible for providing the fetch, query, and web UI for the data provided by the source.


Data is fetched from data sources and loaded into DuckDB. Each data source is responsible for providing fetch functionality (see src/github/ for an example).

The raw data fetched from a source and the raw JSON payload is saved to the wallowa_raw_data table.

The server can be configured to run the fetcher in the background on a time interval.

Web UI and server


The server should not be exposed directly to the Internet since it has not been hardened for that environment. Run a proxy in front of the server if you choose to expose it to the Internet.

Server-side templates are rendered for the web UI. Typescript/Javascript are used to render charts and for some user interactions.

The core web functionality and types are found in src/ Each source is responsible for providing it's web UI (see src/github/ for an example).


The CLI provides the commands and options for the command line interface. The code for the CLI can be found in src/ This includes the ability to generate project scaffolding.