smtp-relay

This project replace a previous project aws-smtp-relay

The goal is to have a dynamic SMTP server that can either be used to run a debug SMTP locally that just store received email in a folder Or relay a SMTP protocol to an SES API call (goal of aws-smtp-relay) Or simulate some Incoming capabilities of AWS SES, like mail2s3 or mail2sqs and similar mail2gcpstorage and mail2gcppubsub

Quick Start

Replace aws-smtp-relay

Docker command

docker run -p 10025:10025 loopingz/smtp-relay:latest configs/aws-smtp-relay.jsonc

Configuration file can leverage the published schema

{
  "$schema": "https://raw.githubusercontent.com/loopingz/smtp-relay/main/config.schema.json"
}

Replace main in url by the tag version to get the configuration format of a specific version

Run with a configuration file:

// Replace my previous project aws-smtp-relay
{
  "$schema": "https://raw.githubusercontent.com/loopingz/smtp-relay/main/config.schema.json",
  "flows": {
    "localhost": {
      "filters": [
        // Allow any ip to use the SMTP
        {
          "type": "whitelist",
          "ips": ["regexp:.*"]
        }
      ],
      "outputs": [
        {
          "type": "aws",
          // Send it to SES
          "ses": {}
        }
      ]
    }
  },
  "options": {
    "disableReverseLookup": false,
    // Do not require auth
    "authOptional": true,
    "loggers": [
      {
        "level": "INFO",
        "type": "CONSOLE"
      },
      {
        "level": "INFO",
        "type": "FILE",
        "filepath": "./smtp.log"
      }
    ]
  }
}

SMTP 2 GCP Storage

{
  "flows": {
    "localhost": {
      "filters": [
        // Allow any ip to use the SMTP
        {
          "type": "whitelist",
          "to": ["regexp:.*@mydomain.com"]
        }
      ],
      "outputs": [
        {
          "type": "gcp",
          // Store it in the bucket
          "path": "gs://myemail/",
          // Send a message to the queue containing the bucket url if exist and the metadata of the email
          "pubsub": ""
        }
      ]
    }
  },
  "options": {
    "disableReverseLookup": false,
    // Do not require auth
    "authOptional": true
  }
}

Run it locally for dev

You can just leveraging the Docker image

docker run -p 10025:10025 -v `pwd`/emails:/smtp-relay/received_emails loopingz/smtp-relay:latest ./configs/fake-smtp.jsonc
# With auth
docker run -e SMTP_USERNAME=test -e SMTP_PASSWORD=plain:test -p 10025:10025 -v `pwd`/emails:/smtp-relay/received-emails loopingz/smtp-relay:latest configs/fake-smtp-with-auth.jsonc

Concepts

The smtp server is subdivided with:

• Filters
• Core
• Processors
• Flows

Filters

These components decide to accept or refuse an email.

At each SMTP command step, they can make a decision to refuse or accept an email or not make a decision boolean|undefined

By default, 3 filters exist:

• whitelist: allow emails based on regexp or exact values
• http-auth: proxy the decision on the email to an HTTP endpoint
• static-auth: staticly defined user/password for authentication

Processors

These components receive the email sent after it was accepted by the filters.

There is 4 types:

• aws:
• gcp:
• file:
• mailer:

Flows

A flow is defined within the configuration, it defines the filters and the outputs to apply if the message match the filters

You can have as many flow as you desire within the SMTP server

Core

Manage the coordination of different component and is in charge of capturing the mail stream

Common variables available for replacements

iso8601: date and time in YYYYmmddHHiiss format

timestamp: UNIX timestamp

id: Session id

The following variables are not always available but should be within processors

from: Email of the sender

messageId: Message id

subject: subject of the email

to: list of recipient comma separated

Logs

You can define log configuration with the loggers definition.

We currently support "CONSOLE" or "FILE"

"loggers": [
  {
    "level": "INFO",
    "type": "CONSOLE"
  },
  {
    "level": "INFO",
    "type": "FILE",
    "filepath": "./smtp.log",
    "sizeLimit": 50000000
  }
]

From the library @webda/workout, the loglevel if not defined fallback to the LOG_LEVEL environment variable and then fallback again to INFO

The FILE type have a size limit defined and will increment a number at the end of the filepath if needed. It has a default sizeLimit define by the library.

A format can be defined too

By default the loggers are defined as a single CONSOLE logger. You can disable completely by adding a loggers: [] property

Contributors ✨

Thanks goes to these wonderful people (emoji key):

This project follows the all-contributors specification. Contributions of any kind welcome!