Skip to content

svix/svix-webhooks

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

GitHub tag Build Status Server Security Twitter Follow Join our slack

Docker Pulls NPM Downloads Pypi Downloads

Svix is the enterprise ready webhook service

Svix makes it easy for developers to send webhooks. Developers make one API call, and Svix takes care of deliverability, retries, security, and more. For more information, please refer to the Svix homepage.

PyPI Crates.io NPM version Gem Maven Central (Java) Maven Central (Kotlin) Nuget Packagist Version PkgGoDev

Documentation

You can find general usage documentation at https://docs.svix.com. For complete API documentation with code examples for each endpoint in all of our official client libraries head over to our API documentation site at https://api.svix.com.

Support & Community

To stay up-to-date with new features and improvements be sure to watch our repo!

Watch & Star our repo

Client Library Overview

⚡️ Feature Breakdown ⚡️
Language Officially Supported API Support Webhook Verification Other Notes
Go
Python
Typescript/Javascript
Java Async support planned. (If you use kotlin, checkout our kotlin library for coroutine support.)
Kotlin
Ruby
C# (dotnet)
Rust
PHP 🔜

Running the server

There are multiple ways to get the Svix server up running. Docker is probably the most common one, but you can choose the one that works best for you.

The Svix server is written in Rust 🦀, which means you can compile it into a static library for a variety of targets. Please refer to the building from source section below for more information.

Please refer to the server configuration section below for more information regarding the available settings.

Deployment

Docker

You can use the official Svix Docker image from Docker Hub. You can either use the latest tag, or one of the versioned tags instead.

You can either use the example docker-compose.yml file with docker compose (easiest), docker swarm (advanced), or run the container standalone.

With Docker Compose

This alternative is the easiest because it will also boot up and configure redis and postgresql.

This assumes you have Docker Compose v2 installed.

cd server
docker compose up

Standalone container

Running a standalone container is slightly more advanced, as it requires you to set some environment variables and have them pointing to your redis and postgres instances. You can pass individual environment variables to docker using the -e flag, or just create a file like development.env and use the --env-file flag like in the example below:

docker run \
  --name svix-server \
  -p 8071:8071 \
  --env-file development.env \
  svix/svix-server

Pre-compiled binaries

Pre-compiled binaries are available for released versions in the releases section.

Building from source

The Svix server is written in Rust 🦀 and requires a Rust build environment.

If you already have one, you just need to run cargo build, otherwise, please please refer to the Svix server README for more information about building the server from source.

Runtime dependencies

The server requires the following runtime dependencies to work correctly:

  • A PostgreSQL server - for the storage of events.
  • An optional Redis server version 6.2.0 or higher - for the task queue and cache.

Redis/Valkey Considerations

Persistence

Please note that it's recommended to enable persistence in Redis so that tasks are persisted across Redis server restarts and upgrades.

Eviction Policy

Please ensure that your Redis instances are configured to not evict keys without explicit expire policies set. This means that maxmemory-policy should be set to noeviction or to any of the available volatile- policies. See Redis/Valkey documentation for further information.

Server configuration

There are three ways to configure svix-server: environment vars, .env file, and a configuration file.

Configuration file

You can put a file called config.toml in the current working directory of svix-server and it will automatically pick it up. You can take a look at the example file for more information and a full list of supported settings: config.toml.

Here's a quick example of the most important configurations:

# The JWT secret for authentication - should be secret and securely generated
jwt_secret = "8KjzRXrKkd9YFcNyqLSIY8JwiaCeRc6WK4UkMnSW"

# The DSN for the database. Only postgres is currently supported.
db_dsn = "postgresql://postgres:postgres@pgbouncer/postgres"

# The DSN for redis (can be left empty if not using redis)
redis_dsn = "redis://redis:6379"

# What kind of message queue to use.
queue_type = "redis"

Environment (variables or .env)

Alternatively, you can configure svix-server by setting the equivalent environment variables for each of the supported settings. The environment variables can either be passed directly or by setting them in a .env file.

The environment variables have the name name as the config names, but they are all upper case and are prefixed with SVIX_.

For example, the above example configuration would look like this if it was passed in the env:

# The JWT secret for authentication - should be secret and securely generated
SVIX_JWT_SECRET = "8KjzRXrKkd9YFcNyqLSIY8JwiaCeRc6WK4UkMnSW"

# The DSN for the database. Only postgres is currently supported.
SVIX_DB_DSN = "postgresql://postgres:postgres@pgbouncer/postgres"

# The DSN for redis (can be left empty if not using redis)
SVIX_REDIS_DSN = "redis://redis:6379"

# What kind of message queue to use.
SVIX_QUEUE_TYPE = "redis"

OpenTelemetry

You may send tracing information to the OpenTelemetry Collector which allows forwarding trace events to a number of external applications/services such as DataDog, Jaeger, NewRelic, Prometheus, Sentry, Signoz, and Zipkin.

You can see more in these instructions.

Connection Pool Size

The db_pool_max_size configuration parameter controls the maximum allowed size of the connection pool for PostgreSQL. This value defaults to a max size of 100, but you can potentially increase application performance significantly by increasing this value. You may need to consider Postgres and PGBouncer configuration parameters as well when tuning these parameters.

The redis_pool_max_size parameter controls the maximum size of each Svix instance's Redis connection pool. Its default is 100. Note that only redis queuing leverages connection pooling -- caching is fully asynchronous and so does not otherwise benefit from pooling. Therefore, you probably won't need to tune this parameter.

SSRF Attacks and Internal IP Addresses

To prevent SSRF attacks, message dispatches to internal IP addresses are blocked by default. However we understand that this doesn't meet the needs of every user say, for example, the service can only be accessed internally. To bypass these restrictions, see the whitelist_subnets configuration option, which accepts an array of CIDR-notation subnets to allow messages to be dispatched to.

Webhook signature scheme (symmetric vs asymmetric)

To ensure the security and integrity of messages, Svix signs all webhook messages prior to sending. Svix supports two types of signature schemes: symmetric (pre-shared key) and asymmetric (public key).

Symmetric signatures are significantly faster (~50x for signing, and ~160x for verifying), and are much simpler (which makes verification easier for your customers), though they require the usage of a pre-shared key per endpoint (endpoint secret) in order to work. Asymmetric signatures on the other hand only require sharing a public key with your customers (not secret).

Because of the above, using symmetric keys is both recommended and the Svix default. Using them is documented in the verifying signatures section of the docs.

However, in some scenarios it may be beneficial to use asymmetric signatures, which is why they too are supported. For more information please refer to the asymmetric signatures section below.

Authentication

Use valid JWTs generated with the correct secret as Bearer.

E.g:

Authorization: Bearer <JWT_TOKEN_HERE>

Either generate one using

svix-server jwt generate

Or if you are generating your own, make sure to use org_23rb8YdGqMT0qIzpgGwdXfHirMu as the sub field, and H256 as the algorithm.

Example valid JWT for the secret x (so you can see the structure):

// JWT: eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJpYXQiOjE2NTUxNDA2MzksImV4cCI6MTk3MDUwMDYzOSwibmJmIjoxNjU1MTQwNjM5LCJpc3MiOiJzdml4LXNlcnZlciIsInN1YiI6Im9yZ18yM3JiOFlkR3FNVDBxSXpwZ0d3ZFhmSGlyTXUifQ.USMuIPrqsZTSj3kyWupCzJO9eyQioBzh5alGlvRbrbA
// Structure (when decoded):
{
  "iat": 1655140639,
  "exp": 1970500639,
  "nbf": 1655140639,
  "iss": "svix-server",
  "sub": "org_23rb8YdGqMT0qIzpgGwdXfHirMu"
}

Using a different signing algorithm

As mentioned above, the default algorithm for signing JWTs is HS256. You can select a different algorithm by setting the jwt_algorithm config to one of these supported values: HS384, HS512, RS256, RS384, RS512, or EdDSA.

Operational (incoming) webhooks

Operational webhooks are webhooks that you can subscribe to in order to get notified of important events occurring on the svix-server. The list of supported events is available in the webhooks section of the API reference.

The operational webhooks utilize Svix, and are controlled by a special account service account with the following ID: org_00000000000SvixManagement00.

The first step is to turn it on by setting the operational_webhook_address config to point to your Svix server. The most common value for this setting is http://127.0.0.1:8071, though it may be different based on your specific setup.

The above step enables operational webhooks on this instance, and the next step is to enable it for your specific organization. As mentioned above, operational webhooks use a normal Svix account behind the scenes, so we'll first need to get the authentication token for this account. To do this you should run:

svix-server jwt generate org_00000000000SvixManagement00

This will give you a special JWT to access the operational webhooks account which is different to the normal JWT you use when interacting with Svix. Let's assume for example that the JWT it returned was op_webhook_token_123.

To enable operational webhooks for a specific account we need to first create an application for it in the service account (remember: operational webhooks just use Svix behind the scenes). We'll use the default Svix account as an example: org_23rb8YdGqMT0qIzpgGwdXfHirMu.

curl -X 'POST' \
  'http://localhost:8071/api/v1/app/' \
  -H 'Authorization: Bearer op_webhook_token_123' \
  -H 'Accept: application/json' \
  -H 'Content-Type: application/json' \
  -d '{
        "name": "Operational webhook for default org",
        "uid": "org_23rb8YdGqMT0qIzpgGwdXfHirMu"
    }'

This is it, we now have operational webhooks enabled for the default account. The only thing left is adding an endpoint where the operational webhooks are going to be sent to. For example:

curl -X 'POST' \
  'https://api.eu.svix.com/api/v1/app/org_23rb8YdGqMT0qIzpgGwdXfHirMu/endpoint/' \
  -H 'Authorization: Bearer AUTH_TOKEN' \
  -H 'Accept: application/json' \
  -H 'Content-Type: application/json' \
  -d '{
        "url": "https://operational-webhook-destination.com/webhook/",
        "filterTypes": [
          “endpoint.updated”,
          “endpoint.deleted”
        ],
    }'

Note how we use the org ID of the default account as the app_id (or rather uid in this case), when creating an endpoint.

That's it. You should now have working operational webhooks. If you ever want to create a new endpoint, or modify an existing endpoint, you just need to generate a JWT for the service account, and then use the JWT like you would use any other Svix account.

Asymmetric signatures

As mentioned above, symmetric signatures are recommended. However, please read the following instructions on setting up asymmetric signatures if you have determined that asymmetric signatures are what you need.

Configuring the keys

By default, the Svix server generates symmetric secrets for endpoints, which in turn means messages will be signed with symmetric keys. To change this default, set the default_signature_type config to ed25519 as follows:

default_signature_type = "ed25519"

Additionally, no matter what the default is set to, you can still override it by explicitly setting a key on an endpoint. To set a symmetric key, set the endpoint secret to a secret prefixed with whsec_, such as whsec_51TKyHBy5KFY1Ab98GQ8V60BkWnejkWy. To set an asymmetric key, set the endpoint secret to a valid ed25519 base64 encoded private key prefixed with whsk_ such as: whsk_6Xb/dCcHpPea21PS1N9VY/NZW723CEc77N4rJCubMbfVKIDij2HKpMKkioLlX0dRqSKJp4AJ6p9lMicMFs6Kvg==.

Please note, that the expected private key structure is: whsk_${base64(private_key + public_key)}.

For testing purposes, new asymmetric key pairs can be generated using the following command:

$ svix-server asymmetric-key generate

Secret key: whsk_6Xb/dCcHpPea21PS1N9VY/NZW723CEc77N4rJCubMbfVKIDij2HKpMKkioLlX0dRqSKJp4AJ6p9lMicMFs6Kvg==
Public key: whpk_1SiA4o9hyqTCpIqC5V9HUakiiaeACeqfZTInDBbOir4=

Signature scheme

Svix uses ed25519(m) for signing the webhook messages, and it constructs m the same way as it does for the symmetric signatures.

When verifying the message you should also ensure that the timestamp is recent enough in order to limit the potential of replay attacks as noted in the symmetric verification docs.

Shutting down the server

To support graceful shutdown on the server, all running tasks are finished before shutting down on a SIGINT/SIGTERM. This usually takes less than ten seconds.

Differences to the Svix hosted service

One of our main goals with open sourcing the Svix dispatcher is ease of use. The hosted Svix service, however, is quite complex due to our scale and the infrastructure it requires. This complexity is not useful for the vast majority of people and would make this project much harder to use and much more limited. This is why this code has been adjusted before being released, and some of the features, optimizations, and behaviors supported by the hosted dispatcher are not yet available in this repo. With that being said, other than some known incompatibilities, the internal Svix test suite passes. This means they are already mostly compatible, and we are working hard on bringing them to full feature parity.

Re-driving Redis DLQ

We have an undocumented endpoint for re-driving failed messages that are DLQ'ed. You can do this by calling POST /api/v1/admin/redrive-dlq/.

To monitor the DLQ depth, you should monitor the svix.queue.depth_dlq metric. Any non-zero values indicate that there is data in the DLQ.

Development

Checkout our project specific development guides to get started hacking on Svix!

Contributing

Contributions are what makes the open source world go round! All contributions are very much welcomed and are greatly appreciated.

Please refer to the contribution guide for information on how to contribute.

A quick how to for contribution:

  1. Fork the project
  2. Create your feature branch (git checkout -b feature/some-feature)
  3. Make your changes
  4. Commit your changes (git commit -m 'Implement an amazing feature.')
  5. Push to the branch (git push origin feature/some-feature)
  6. Open a pull request

License

Distributed under the MIT License. See LICENSE for more information.

Sending guides

Here is a list of guides for sending webhooks with Svix:

Backed By

Backed By YC & Aleph