Research
Security News
Malicious npm Package Targets Solana Developers and Hijacks Funds
A malicious npm package targets Solana developers, rerouting funds in 2% of transactions to a hardcoded address.
github.com/svix/svix-webhooks
Svix - Webhooks as a 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.
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.
To stay up-to-date with new features and improvements be sure to watch our repo!
⚡️ 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 | ✅ | 🔜 | ✅ |
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.
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.
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
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 are available for released versions in the releases section.
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.
The server requires the following runtime dependencies to work correctly:
Please note that it's recommended to enable persistence in Redis so that tasks are persisted across Redis server restarts and upgrades.
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.
There are three ways to configure svix-server
: environment vars, .env
file, and a 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"
.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"
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.
There are two configuration variables db_pool_max_size
and redis_pool_max_size
which control the maximum allowed size of the connection pool for PostgreSQL and Redis respectively.
They default to a max size of 20, but higher values can significantly increase performance if your database can handle it.
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.
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.
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"
}
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 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.
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.
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=
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.
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.
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.
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.
Checkout our project specific development guides to get started hacking on Svix!
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:
git checkout -b feature/some-feature
)git commit -m 'Implement an amazing feature.'
)git push origin feature/some-feature
)Distributed under the MIT License. See LICENSE for more information.
Here is a list of guides for sending webhooks with Svix:
FAQs
Unknown package
Did you know?
Socket for GitHub automatically highlights issues in each pull request and monitors the health of all your open source dependencies. Discover the contents of your packages and block harmful activity before you install or update your dependencies.
Research
Security News
A malicious npm package targets Solana developers, rerouting funds in 2% of transactions to a hardcoded address.
Security News
Research
Socket researchers have discovered malicious npm packages targeting crypto developers, stealing credentials and wallet data using spyware delivered through typosquats of popular cryptographic libraries.
Security News
Socket's package search now displays weekly downloads for npm packages, helping developers quickly assess popularity and make more informed decisions.