Expand description

Easy boilerplate utilities for Rust http services which use async-std, Tide, Surf, and friends.

Allows for service setup with feature-configured built-ins for maximum service consistency with low developer overhead, and for easily integration testing the service without using a live network.

Scroll to the bottom for API Reference


use std::sync::Arc;

use tide::{Request, Route};

struct AppState {
    greeting: &'static str,

type AppRequest = Request<Arc<AppState>>;

async fn setup_app_state() -> preroll::SetupResult<AppState> {
    Ok(AppState {
        greeting: "Hello World!",

fn setup_routes(mut server: Route<'_, Arc<AppState>>) {
        .get(|req: AppRequest| async move {

// The "magic" happens here!
preroll::main!("hello-world", setup_app_state, setup_routes);


  • Boilerplate main setup via preroll::main!, with optional features automatically configured.
  • A preroll::prelude::*; with all extension traits.
  • Response logging with many details.
  • Automatic JSON responses for errors in the form of JsonError.
  • Test utils with easy mock client setup.

Optional features

Add-on features must be enabled via cargo features, e.g.

version = "0.5"
features = ["honeycomb", "postgres"]

List of optional add-on features:

  • "honeycomb": Enables tracing to honeycomb.io.
    • Env variable HONEYCOMBIO_WRITE_KEY (required).
    • Env variable TRACELEVEL, sets the tracing level filter, defaults to info.
    • Writes to a dataset named {service_name}-{environment}.
      • service_name is from preroll::main!("service_name", ...).
      • environment is from ENVIRONMENT, or defaults to "development".
  • "lambda-http": Changes the HTTP listener to connect to an AWS Lambda execution environment.
    • Is no longer reachable as a regular http server, but accepts http lambda requests as if it were one.
    • Some environment variables, such as PORT, are disregarded.
    • If the "honeycomb" feature is enabled, trace events are written to stdout, and must be collected via a layer provided by Honeycomb. See: https://docs.honeycomb.io/getting-data-in/integrations/aws/aws-lambda/
  • "postgres": Enables a postgres connection pool with transactions.
    • Env variable PGURL, which should be a properly formatted postgres:// database url.
      • Defaults to "postgres://localhost/{service_name}" (default postgres port).
      • service_name is from preroll::main!("service_name", ...).
    • Env variable PGMAXCONNECTIONS, default 5 connections.
    • Env variable PGMAXLIFETIME, default 30 (minutes).
    • Enables PostgresRequestExt and test_utils::create_client_and_postgres.

List of other optional features:

  • "panic-on-error": Makes the response logger panic on error rather than log.
    • Do not use in production. Prevents --release compilation.

General Environment Settings

The following environment variables are read during preroll::main!:

  • ENVIRONMENT: If this starts with prod, load the production-mode JSON logger, avoid .env.
  • FORCE_DOTENV: Override production-mode, force-load environment from .env.
  • HOST: Sets the hostname that this service will listen on. Defaults to "".
  • LOGLEVEL: Set the logger’s level filter, defaults to info in production-mode, debug in development-mode.
  • PORT: Sets the port that this service will listen on. Defaults to 8080.


This crate is intentionally somewhat prescriptive in how it templates a service and the interaction with add-on features such as Postgres (via SQLx).


Auto-import of all preroll extension traits.

Utilities for setting up mock clients and test servers with similar features to preroll::main!.

Miscellaneous utilities.


Begin here. A macro which constructs the equivalent of an async fn main() {}.


The structure of an error as formatted by preroll’s error handling middleware.

Variadic-argument route versioning is implemented via this struct for From<T> with Single-argument, Tuple, and Vec types.

Type Definitions

The result type which is expected from functions passed to preroll::main!.