wheniwork/ld-relay repository overview

⁠LaunchDarkly Relay Proxy

⁠What is it?

The LaunchDarkly Relay Proxy establishes a connection to the LaunchDarkly streaming API, then proxies that stream connection to multiple clients.

The relay proxy lets a number of servers connect to a local stream instead of making a large number of outbound connections to stream.launchdarkly.com.

The relay proxy can be configured to proxy multiple environment streams, even across multiple projects. It can also be used as a local proxy that forwards events to events.launchdarkly.com. This can be useful if you are load balancing LDR instances behind a proxy that times out HTTP connections (e.g. Elastic Load Balancers).

⁠When should it be used?

In most cases, the relay proxy is not required. However, there are some specific scenarios where we recommend deploying the proxy to improve performance and reliability:

1. PHP-- PHP has a shared-nothing architecture that prevents the normal LaunchDarkly streaming API connection from being re-used across requests. While we do have a supported deployment mode for PHP that does not require the relay proxy, we strongly recommend using the proxy in daemon mode (see below) if you are using PHP in a high-throughput setting. This will offload the task of receiving feature flag updates to the relay proxy. We also recommend using the relay to forward events to events.launchdarkly.com, and configuring the PHP client to send events to the relay synchronously. This eliminates the curl / fork method that the PHP SDK uses by default to send events back to LaunchDarkly asynchronously.

2. Reducing outbound connections to LaunchDarkly-- at scale (thousands or tens of thousands of servers), the number of outbound persistent connections to LaunchDarkly's streaming API can be problematic for some proxies and firewalls. With the relay proxy in place in proxy mode, your servers can connect directly to hosts within your own datacenter instead of connecting directly to LaunchDarkly's streaming API. On an appropriately spec'd machine, each relay proxy can handle tens of thousands of concurrent connections, so the number of outbound connections to the LaunchDarkly streaming API can be reduced dramatically.

3. Reducing redundant update traffic to Redis-- if you are using Redis as a shared persistence option for feature flags, and have a large number of servers (thousands or tens of thousands) connected to LaunchDarkly, each server will attempt to update Redis when a flag update happens. This pattern is safe but inefficient. By deploying the relay proxy in daemon mode, and setting your LaunchDarkly SDKs to daemon mode, you can delegate flag updates to a small number of relay proxy instances and reduce the number of redundant update calls to Redis.

⁠Quick setup

1. Make sure go 1.6+ is installed. Follow the instructions provided in Go's documentation⁠

2. Build and install the binary in your $GOPATH:

go get -u gopkg.in/launchdarkly/ld-relay.v5/...

1. Create LD Relay configuration file. Create a new filed called ld-relay.conf containing the text:

[main]
    streamUri = "https://stream.launchdarkly.com"
    baseUri = "https://app.launchdarkly.com"
[environment "<NAME-OF-YOUR-ENVIRONMENT>]
    sdkKey = "<SDK-KEY-FOR-YOUR-ENVIRONMENT>

1. Run the binary by entering the following command in your terminal:

$GOPATH/bin/ld-relay --config ./ld-relay.conf

1. Validate functionality. This can be done two different ways.

• Update your SDKs configuration. When initializing the SDK, set configuration attribute streamUri to the host and port of your relay proxy instance. Note: Your SDK must not be set to polling mode. You must leave streaming enabled to use LD-Relay.
• Manually curl the relay. Run the following command in your terminal:

curl -X REPORT localhost:8030/sdk/eval/user -H "Authorization: YOUR_SDK_KEY" -H "Content-Type: application/json" -d '{"key": "a00ceb", "email":"[email protected]"}'

⁠Command-line arguments

⁠Configuration file format

LD Relay uses INI-style configuration files. You can read more about the syntax here⁠.

There are four primary section types: Main, Events, Redis, and Environments. In addition to the main section types, there are three supported sections for configuring exporters for exporting metrics and route traces: Datadog, Stackdriver, and Prometheus.

⁠[main]

⁠[events]

⁠[redis]

⁠[environment]

⁠[datadog]

⁠[stackdriver]

⁠[prometheus]

[main]
    streamUri = "https://stream.launchdarkly.com"
    baseUri = "https://app.launchdarkly.com"
    exitOnError = true
    heartbeatIntervalSecs = 15

[environment "Spree Project Production"]
    sdkKey = "SPREE_PROD_API_KEY"

[environment "Spree Project Test"]
    sdkKey = "SPREE_TEST_API_KEY"

[environment "Shopnify Project Production"]
    sdkKey = "SHOPNIFY_PROD_API_KEY"

[environment "Shopnify Project Test"]
    sdkKey = "SHOPNIFY_TEST_API_KEY"

⁠Mobile and client-side flag evaluation

LDR may be optionally configured with a mobile SDK key, and/or an environment ID to enable flag evaluation support for mobile and client-side LaunchDarkly SDKs (Android, iOS, and JavaScript).

[environment "Spree Mobile Production"]
    sdkKey = "SPREE_MOBILE_PROD_API_KEY"
    mobileKey = "SPREE_MOBILE_PROD_MOBILE_KEY"

[environment "Spree Webapp Production"]
    sdkKey = "SPREE_WEB_PROD_API_KEY"
    envId = "SPREE_WEB_PROD_ENV_ID"
    allowedOrigin = "http://example.org"
    allowedOrigin = "http://another_example.net"

Once a mobile key or environment ID has been configured, you may set the baseUri parameter to the host and port of your relay proxy instance in your mobile/client-side SDKs. If you are exposing any of the client-side relay endpoints externally, https should be configured with a TLS termination proxy.

⁠Event forwarding

LDR can also be used to forward events to events.launchdarkly.com. When enabled, the relay will buffer and forward events posted to /bulk to https://events.launchdarkly.com/bulk. The primary use case for this is PHP environments, where the performance of a local proxy makes it possible to synchronously flush analytics events. To set up event forwarding, add an events section to your configuration file:

[events]
    eventsUri = "https://events.launchdarkly.com"
    sendEvents = true
    flushIntervalSecs = 5
    samplingInterval = 0
    capacity = 10000
    inlineUsers = false

This configuration will buffer events for all environments specified in the configuration file. The events will be flushed every flushIntervalSecs. To point our SDKs to the relay for event forwarding, set the eventsUri in the SDK to the host and port of your relay instance (or preferably, the host and port of a load balancer fronting your relay instances). Setting inlineUsers to true preserves full user details in every event (the default is to send them only once per user in an "index" event).

⁠Redis storage

You can configure LDR nodes to persist feature flag settings in Redis. This provides durability in case of (e.g.) a temporary network partition that prevents LDR from communicating with LaunchDarkly's servers.

[redis]
    host = "localhost"
    port = 6379
    localTtl = 30000

⁠Relay proxy mode

LDR is typically deployed in relay proxy mode. In this mode, several LDR instances are deployed in a high-availability configuration behind a load balancer. LDR nodes do not need to communicate with each other, and there is no master or cluster. This makes it easy to scale LDR horizontally by deploying more nodes behind the load balancer.

⁠Daemon mode

Optionally, you can configure our SDKs to communicate directly to the Redis store. If you go this route, there is no need to put a load balancer in front of LDR-- we call this daemon mode. This is the preferred way to use LaunchDarkly with PHP (as there's no way to maintain persistent stream connections in PHP).

To set up LDR in this mode, provide a redis host and port, and supply a Redis key prefix for each environment in your configuration file:

[redis]
    host = "localhost"
    port = 6379
    localTtl = 30000

[main]
    ...

[environment "Spree Project Production"]
    prefix = "ld:spree:production"
    sdkKey = "SPREE_PROD_API_KEY"

[environment "Spree Project Test"]
    prefix = "ld:spree:test"
    sdkKey = "SPREE_TEST_API_KEY"

You can also configure an in-memory cache for the relay to use so that connections do not always hit redis. To do this, set the localTtl parameter in your redis configuration section to a number (in milliseconds).

If you're not using a load balancer in front of LDR, you can configure your SDKs to connect to Redis directly by setting use_ldd mode to true in your SDK, and connecting to Redis with the same host and port in your SDK configuration.

⁠Flag evaluation endpoints

If you're building an SDK for a language which isn't officially supported by LaunchDarkly, or would like to evaluate feature flags internally without an SDK instance, the relay provides endpoints for evaluating all feature flags for a given user. These endpoints support the GET and REPORT http verbs to pass in users either as base64url encoded path parameters, or in the request body, respectively.

Example cURL requests (default local URI and port):

curl -X GET -H "Authorization: YOUR_SDK_KEY" localhost:8030/sdk/eval/users/eyJrZXkiOiAiYTAwY2ViIn0=

curl -X REPORT localhost:8030/sdk/eval/user -H "Authorization: YOUR_SDK_KEY" -H "Content-Type: application/json" -d '{"key": "a00ceb", "email":"[email protected]"}'

⁠Performance, scaling, and operations

We have done extensive load tests on the relay proxy in AWS / EC2. We have also collected a substantial amount of data based on real-world customer use. Based on our experience, we have several recommendations on how to best deploy, operate, and scale the relay proxy:

• Networking performance is paramount. Memory and CPU are not as critical. The relay proxy should be deployed on boxes with good networking performance. On EC2, we recommend using an instance with Moderate to High networking performance⁠ such as m4.xlarge. On an m4.xlarge instance, a single relay proxy node can easily manage 20,000 concurrent connections.

• If using an Elastic Load Balancer in front of the relay proxy, you may need to pre-warm⁠ the load balancer whenever connections to the relay proxy are cycled. This might happen when you deploy a large number of new servers that connect to the proxy, or upgrade the relay proxy itself.

⁠Proxied endpoints

The table below describes the endpoints proxied by the LD relay. In this table:

• user is the base64 representation of a user JSON object (e.g. *"key": "user1"* => eyJrZXkiOiAidXNlcjEifQ==).
• clientId is the 32-hexdigit Client-side ID (e.g. 6488674dc2ea1d6673731ba2)

⁠Exporting metrics and traces

The relay may be configured to export statistics and route traces to Datadog, Stackdriver, and Prometheus. See the configuration section⁠ for configuration instructions.

The following metrics are supported:

• connections: The number of current proxied streaming connections.
• newconnections: The number of streaming connections created.
• requests: Number of requests received.

Metrics can be filtered by the following tags:

• platformCategoryTagKey: The platform a metric was generated by (e.g. server, browser, or client-side).
• env: The name of the LaunchDarkly environment.
• route: The request route.
• method: The http method used for the request.
• userAgent: The user agent used to make the request, typically a LaunchDarkly SDK version. Example: "Node/3.4.0"

Note: Traces for stream connections will trace until the connection is closed.

⁠Docker

Using docker is not required, but if you perfer using a docker container we provide a docker entrypoint to make this as easy as possible.

To build the ld-relay container:

$ docker build -t ld-relay .

⁠Docker environment variables

The docker entrypoint uses environment variables to configured the dockerized LD Relay instance.

⁠Docker examples

To run a single environment, without Redis:

$ docker run --name ld-relay -e LD_ENV_test="sdk-test-sdkKey" ld-relay

To run multiple environments, without Redis:

$ docker run --name ld-relay -e LD_ENV_test="sdk-test-sdkKey" -e LD_ENV_prod="sdk-prod-sdkKey" ld-relay

To run a single environment, with Redis:

$ docker run --name redis redis:alpine
$ docker run --name ld-relay --link redis:redis -e USE_REDIS=1 -e LD_ENV_test="sdk-test-sdkKey" ld-relay

To run multiple environment, with Redis:

$ docker run --name redis redis:alpine
$ docker run --name ld-relay --link redis:redis -e USE_REDIS=1 -e LD_ENV_test="sdk-test-sdkKey" -e LD_PREFIX_test="ld:default:test" -e LD_ENV_prod="sdk-prod-sdkKey" -e LD_PREFIX_prod="ld:default:prod" ld-relay

⁠Windows

To register ld-relay as a service, run a command prompt as Administrator

$ sc create ld-relay DisplayName="LaunchDarkly Relay Proxy" start="auto" binPath="C:\path\to\ld-relay.exe -config C:\path\to\ld-relay.conf"

⁠Integrating LD Relay into your own application

You can also use relay to handle endpoints in your own application if you don't want to use the default ld-relay application. Below is an example using Gorilla⁠ of how you might instantiate a relay inside your web server beneath a path called "/relay":

router := mux.NewRouter()
configFileName := "path/to/my-config-file"
cfg := relay.DefaultConfig
if err := relay.LoadConfigFile(&cfg, configFileName); err != nil {
  log.Fatalf("Error loading config file: %s", err)
}
r, err := relay.NewRelay(cfg, relay.DefaultClientFactory)
if err != nil {
  log.Fatalf("Error creating relay: %s", err)
}
router.PathPrefix("/relay").Handler(r)