Rust Rest API Stack with User Management, Kafka Message Publishing, S3 uploads/downloads, and Prometheus for Monitoring

A secure-by-default Rest API using hyper, tokio, bb8, kafka-threadpool, postgres, and prometheus for monitoring.

Features

1. User management and authentication stored in postgres
2. Async s3 uploading and downloading (to/from local files or to/from memory)
3. Decoupled, async kafka threadpool that uses environment variables to connect to a kafka cluster with client mtls for authentication and encryption in transit
4. Async publishing for all successful user events to a kafka topic (topic default: user.events) and partition key (key default: user-{user.id})
5. Async kafka messaging for one-off messages using custom kafka topic(s), partition key(s) and custom header(s).

Overview

User

• User password reset and user email change support using one-time-use tokens that are stored in postgres.
• Users can upload and manage files stored on AWS S3 (assuming valid credentials are loaded outside this rust project).
• User passwords are hashed using argon2.

Auth

• User authentication enabled by default
• Default JWT signing keys included with documentation for building new keys as needed.

Database

• The rest api server utilizes postgres with a bb8 client threadpool.
• The postgres database requires each client connection include the postgres tls certificate authority file for encrypting data in-transit.
• Includes pg4admin for database management in a browser (deployed with podman compose).

TLS Encryption

• Includes a tls asset generator tool (./tls/create-tls-assets.sh) for building self-signed tls assets including your own private Certificate Authority (CA).

Ingress

Component	Status
Rest API Server	Listening for encrypted client connections on tcp port 3000
Postgres	Listening for encrypted client connections on tcp port 5432 (tls Certificate Authority required)
pgAdmin	Listening for encrypted HTTP client connections on tcp port 5433

Getting Started

Clone the repo

git clone https://github.com/jay-johnson/restapi
cd restapi

Generate TLS Assets and a Private Certificate Authority (CA) using CFSSL

Generate new tls assets under the ./tls directory with these commands:

cd tls
./create-tls-assets.sh
cd ..

Please refer to the Generating TLS Assets with CFSSL for more information.

Generate JWT Private and Public Signing Keys

Generate new signing JWT keys under the ./jwt directory with these commands:

cd jwt
./recreate-jwt.sh
cd ..

Please refer to the How to build JWT private and public keys for the jsonwebtokens crate doc for more information.

Deploy Postgres and pgAdmin using Podman

Please refer to the Build and Deploy a Secured Postgres backend doc for more information.

Build API Server

cargo build --example server

Run API Server

export RUST_BACKTRACE=1 &&export RUST_LOG=info,kafka_threadpool=info && ./target/debug/examples/server

Environment Variables

Rest API

Environment Variable	Default
SERVER_NAME_API	api
SERVER_NAME_LABEL	rust-restapi
API_ENDPOINT	0.0.0.0:3000
API_TLS_DIR	./tls/api
API_TLS_CA	./tls/ca/ca.pem
API_TLS_CERT	./tls/api/server.pem
API_TLS_KEY	./tls/api/server-key.pem

User Email Verification

Environment Variable	Default
USER_EMAIL_VERIFICATION_REQUIRED	"0"
USER_EMAIL_VERIFICATION_ENABLED	"1"
USER_EMAIL_VERIFICATION_EXP_IN_SECONDS	"2592000"

User One-Time-Use Token Expiration for Password Recovery

Environment Variable	Default
USER_OTP_EXP_IN_SECONDS	"2592000"

Postgres Database

Environment Variable	Default
POSTGRES_USERNAME	datawriter
POSTGRES_PASSWORD	"123321"
POSTGRES_ENDPOINT	0.0.0.0:5432
POSTGRES_TLS_DIR	./tls/postgres
POSTGRES_TLS_CA	./tls/ca/ca.pem
POSTGRES_TLS_CERT	./tls/postgres/client.pem
POSTGRES_TLS_KEY	./tls/postgres/client-key.pem
POSTGRES_DB_CONN_TYPE	postgresql

Kafka Cluster

Please refer to the kafka_threadpool docs for more information.

Environment Variable	Purpose / Value
KAFKA_PUBLISH_EVENTS	if set to true or 1 publish all user events to kafka
KAFKA_ENABLED	toggle the kafka_threadpool on with: true or 1 anything else disables the threadpool
KAFKA_LOG_LABEL	tracking label that shows up in all crate logs
KAFKA_BROKERS	comma-delimited list of brokers (host1:port,host2:port,host3:port)
KAFKA_TOPICS	comma-delimited list of supported topics
KAFKA_PUBLISH_RETRY_INTERVAL_SEC	number of seconds to sleep before each publish retry
KAFKA_PUBLISH_IDLE_INTERVAL_SEC	number of seconds to sleep if there are no message to process
KAFKA_NUM_THREADS	number of threads for the threadpool
KAFKA_TLS_CLIENT_KEY	optional - path to the kafka mTLS key (./tls/kafka-cluster-0/client-key.pem)
KAFKA_TLS_CLIENT_CERT	optional - path to the kafka mTLS certificate (./tls/kafka-cluster-0/client.pem)
KAFKA_TLS_CLIENT_CA	optional - path to the kafka mTLS certificate authority (CA) (./tls/ca/ca.pem)
KAFKA_METADATA_COUNT_MSG_OFFSETS	optional - set to anything but true to bypass counting the offsets

Sample kafka.env file

# enable the clusterexport KAFKA_ENABLED=1
export KAFKA_LOG_LABEL="ktp"export KAFKA_BROKERS="host1:port,host2:port,host3:port"export KAFKA_TOPICS="testing"export KAFKA_PUBLISH_RETRY_INTERVAL_SEC="1.0"export KAFKA_NUM_THREADS="5"export KAFKA_TLS_CLIENT_CA="./tls/ca/ca.pem"export KAFKA_TLS_CLIENT_CERT="./tls/kafka-cluster-0/client.pem"export KAFKA_TLS_CLIENT_KEY="./tls/kafka-cluster-0/client-key.pem"# the KafkaPublisher can count the offsets for each topic with "true" or "1"export KAFKA_METADATA_COUNT_MSG_OFFSETS="true"

S3

Environment Variable	Default
S3_DATA_BUCKET	YOUR_BUCKET
S3_DATA_PREFIX	/rust-restapi/tests
S3_STORAGE_CLASS	STANDARD
S3_DATA_UPLOAD_TO_S3	"0"

JWT

Environment Variable	Default
TOKEN_EXPIRATION_SECONDS_INTO_FUTURE	"2592000"
TOKEN_ORG	example.org
TOKEN_HEADER	Bearer
TOKEN_ALGO_PRIVATE_KEY	./jwt/private-key-pkcs8.pem
TOKEN_ALGO_PUBLIC_KEY	./jwt/public-key.pem
SERVER_PKI_DIR_JWT	./jwt
SERVER_PASSWORD_SALT	78197b60-c950-4339-a52c-053165a04764

Rust

Environment Variable	Default
RUST_BACKTRACE	"1"
RUST_LOG	info

Debug

Environment Variable	Default
DEBUG	"1"

Docker Builds

Build Base Image

This will build an initial base image using podman. Note: this base image will not work on a different cpu chipset because the openssl libraries are compiled within the image for this base image.

./build-base.sh

Build Derived Image

By reusing the base image, this derived image only needs to recompile the server. With minimal code changes, this is a much faster build than the base image build.

./build-derived.sh

Kubernetes

Start Kafka

If you do not have a running Kafka cluster, you can deploy your own with:

https://github.com/jay-johnson/rust-with-strimzi-kafka-and-tls

Helm Chart

Deploy TLS Assets into Kubernetes

This command will deploy all jwt keys, tls assets and credentials into the dev namespace:

./deploy-kubernetes-assets.sh -e dev

Deploy the Rust Rest API into Kubernetes

Please refer to the Deploying the Rust Rest API helm chart into kubernetes guide for deploying the example helm chart into a kubernetes cluster.

By default this uses the jayjohnson/rust-restapi container image

helm upgrade --install -n dev dev-api ./charts/rust-restapi -f ./charts/rust-restapi/values.yaml

Monitoring

Prometheus

This section assumes you have a working prometheus instance already running inside kubernetes. Below is the Prometheus scrape_config to monitor the rest api deployment replica(s) within kubernetes. Note this config also assumes the api chart is running in the dev namespace:

scrape_configs:
- job_name: rust-restapiscrape_interval: 10sscrape_timeout: 5smetrics_path: /metricsscheme: httpstls_config:
insecure_skip_verify: truestatic_configs:
- targets:
- dev-api.dev.svc.cluster.local:3000

Supported APIs

Here are the supported json contracts for each Request and Response based off the url. Each client request is handled by the ./src/handle_requests.rs module and returned as a response back to the client (serialization using serde_json)

User APIs

Create User

Create a single users record for the new user

• URL path: /user
• Method: POST
• Handler: create_user
• Request: ApiReqUserCreate
• Response: ApiResUserCreate

Update User

Update supported users fields (including change user email and password)

• URL path: /user
• Method: PUT
• Handler: update_user
• Request: ApiReqUserUpdate
• Response: ApiResUserUpdate

Get User

Get a single user by users.id - by default, a user can only get their own account details

• URL path: /user/USERID
• Method: GET
• Handler: get_user
• Request: ApiReqUserGet
• Response: ApiResUserGet

Delete User

Delete a single users record (note: this does not delete the db record, just sets the users.state to inactive 1)

• URL path: /user
• Method: DELETE
• Handler: delete_user
• Request: ApiReqUserDelete
• Response: ApiResUserDelete

Search Users in the db

Search for matching users records in the db

• URL path: /user/search
• Method: POST
• Handler: search_users
• Request: ApiReqUserSearch
• Response: ApiResUserSearch

Create One-Time-Use Password Reset Token (OTP)

Create a one-time-use password reset token that allows a user to change their users.password value by presenting the token

• URL path: /user/password/reset
• Method: POST
• Handler: create_otp
• Request: ApiReqUserCreateOtp
• Response: ApiResUserCreateOtp

Consume a One-Time-Use Password Reset Token (OTP)

Consume a one-time-use password and change the user's users.password value to the new argon2-hashed password

• URL path: /user/password/change
• Method: POST
• Handler: consume_user_otp
• Request: ApiReqUserConsumeOtp
• Response: ApiResUserConsumeOtp

Verify a User's email

Consume a one-time-use verification token and change the user's users.verified value verified (1)

• URL path: /user/verify
• Method: GET
• Handler: verify_user
• Request: ApiReqUserVerify
• Response: ApiResUserVerify

User S3 APIs

Upload a file asynchronously to AWS S3 and store a tracking record in the db

Upload a local file on disk to AWS S3 asynchronously and store a tracking record in the users_data table. The documentation refers to this as a user data or user data file record.

• URL path: /user/data
• Method: POST
• Handler: upload_user_data
• Request: ApiReqUserUploadData
• Response: ApiResUserUploadData

Update an existing user data file record for a file stored in AWS S3

Update the users_data tracking record for a file that exists in AWS S3

• URL path: /user/data
• Method: PUT
• Handler: update_user_data
• Request: ApiReqUserUpdateData
• Response: ApiResUserUpdateData

Search for existing user data files from the db

Search for matching records in the users_data db based off the request's values

• URL path: /user/data/search
• Method: POST
• Handler: search_user_data
• Request: ApiReqUserSearchData
• Response: ApiResUserSearchData

User Authentication APIs

User Login

Log the user in and get a json web token (jwt) back for authentication on subsequent client requests

• URL path: /login
• Method: POST
• Handler: login
• Request: ApiReqUserLogin
• Response: ApiResUserLogin

Integration Tests

This project focused on integration tests for v1 instead of only rust tests (specifically everything has been tested with curl):

Please refer to the Integration Tests Using curl Guide

Podman Image Push

cur_tag=$(cat Cargo.toml | grep version | head -1 | sed -e 's/"//g'| awk '{print $NF}')
podman tag IMAGE_ID "docker://docker.io/jayjohnson/rust-restapi:${cur_tag}"
podman tag IMAGE_ID "docker://docker.io/jayjohnson/rust-restapi:latest"
podman push "docker.io/jayjohnson/rust-restapi:${cur_tag}"
podman push "docker.io/jayjohnson/rust-restapi:latest"

Build Docs

cargo doc --example server

Start Development Server

source ./env/api.env &&source ./env/kafka.env &&source ./env/postgres.env && cargo build --example server &&export RUST_BACKTRACE=1 &&export RUST_LOG=info,kafka_threadpool=info,rdkafka=error && ./target/debug/examples/server