Rate limiter

For more information about the feature, see Rate limit.

What data gets stored link

The rate limiter stores information for limiting requests in the backing database, including the following information. For an example of how to check the data that gets stored, see Review data in Redis.

• The generated unique key for the request
• The associated count against the rate limit

The configuration data of the Gloo custom resources are stored by the management server in its backing database.

Backing storage options link

Review the following options for setting up backing storage with Redis®* ¹ instances.

Shared or separate instances link

In Gloo Mesh Gateway, many components use Redis to store data. Several of these components must share the same instance in order to function properly, as follows:

• Management server, UI, and insights engine (which is also used as part of the OTel pipeline).
• External auth service and developer portal server.
• Rate limiter, which is often shared with the external auth service and portal server, but does not have to be.

If you enable the default, Solo-provided, local Redis options, these Redis instances are created separately.

If you bring your own Redis, you can choose to share the Redis instances or not. For internal compliance or security purposes, you might want separate Redis instances for each of these use cases. As such, the following guide assumes a separate deployment, and uses separate secrets. However, you can also share a single external Redis instance and update the names accordingly.

Default local Redis link

When you install Gloo Mesh Gateway, a local Redis instance redis is set up in the add-ons namespace of the workload cluster. The rate limiting server automatically reads and writes to this Redis instance.

1. When you install or upgrade Gloo Mesh Gateway, include the following Helm values to enable the local Redis instance.

     
   rateLimiter:
     enabled: true
     redis:
       # Enable the creation of the local Redis deployment and service.
       enabled: true
     

2. Optional: Further configure the default Redis deployment values, such as to modify the image pull policy. For more information, check the Helm reference for the redis component.

3. Verify that data is stored in the local Redis instance.

Bring your own Redis link

Instead of using the built-in local Redis instance, you can achieve higher availability, disaster recovery, and enhanced control and security by bringing your own Redis cluster.

When you install Gloo Mesh Gateway, you can choose to use add-ons such as rate limiter, external auth, or portal (for Gloo Mesh Gateway). These add-ons often share the same backing Redis database. Even if you want to use the same Redis instance for all of these add-ons, you must configure the Redis details for each add-on separately in the Helm installation. This guide shows how to configure the details for the rate limiter.

Step 1: Choose a Redis provider link

You can choose between a local or externally hosted Redis instance.

• For a local instance: Install the Redis instance in the same cluster as the component that the Redis backs. For example, for the Gloo management server, install the Redis instance in the management cluster. For an add-on, install the Redis instance in the same workload as the rate limiter, external auth service, or portal server.
• For an external cloud instance: Create an instance outside your cluster environment, such as Redis Enterprise or a cloud provider service like AWS ElastiCache.

Step 2: Configure your Redis instance link

To create or use an existing Redis instance, make sure that you can connect to your Redis instance from the Gloo management cluster. These steps can vary by provider, but the following guidelines are common.

• Local instances: The Gloo management server and Redis instances must be able to communicate. For example, you might deploy them in the same Kubernetes namespace.
• Cloud instances: Often, you must configure networking settings such as the following:
  • Create the Kubernetes cluster and the Redis instance in the same virtual private network (VPC).
  • Create the Kubernetes cluster and the Redis instance in the same security group.
  • Make sure that the security group allows inbound and outbound connections between the cluster and Redis instance.
  • If the cluster is behind a firewall, you might need to add the public IP address of the Redis cluster to your allowlist.
• Provider-specific considerations: Continue reading for requirements per Cloud provider.
  • AWS Elasticache

AWS ElastiCache link

When setting up your AWS ElastiCache instance, consider the following deployment options:

• Use a high availability (HA) setup consisting of at least 1 primary and 1 replica instance.
• Use non-clustered mode. Cluster mode is currently not supported.
• Set up multiple availability zones (multi-AZ) with auto-failover enabled.
• Set the maxmemory-policy directive to noeviction. For more information about eviction policies, see the Redis docs.
• Use node types with TLS Offloading and Enhanced I/O Multiplexing.
• Although optional, enabling TLS is recommended.
• Monitor performance metrics including CPU utilization, memory usage, network throughput, and cache hit ratio.
• Based on the size of your environment, increase the type of node instances. For more information, see Size and memory system requirements.

Step 3: Get your Gloo environment details link

You can configure Redis during an initial installation or by performing an upgrade. Either way, you can configure the Redis details in your Helm values file.

Installation link

If you have not installed Gloo Mesh Gateway yet, prepare details for your installation.

1. Decide the Helm release namespace to use. The example uses the names suggested throughout the documentation, gloo-platform and gloo-mesh.

     export GLOO_MESH_RELEASE_NAMESPACE=gloo-mesh
     

2. Create a Helm values file for your installation, such as gloo-single.yaml for a single cluster or mgmt-plane.yaml and data-plane.yaml for a multicluster environment.

Upgrade link

If you are upgrading Gloo Mesh Gateway to set up Redis, get the details of your installation.

1. Save the details of your Gloo Mesh Gateway installation, including the Helm chart namespace for the component that you want to set up the Redis instance for. In the following example, the release name is gloo-platform and the namespace is gloo-mesh, but you can update your values accordingly.

     export GLOO_MESH_RELEASE_NAMESPACE=gloo-mesh
     

   To get the values for your own release, run the following Helm command.

     helm list --all-namespaces
     

     NAME                 	NAMESPACE       	REVISION	UPDATED                             	STATUS  	CHART
   gloo-platform        	gloo-mesh       	1      	  2023-09-01 13:26:56.061102 -0400 EDT	deployed	gloo-platform-2.4.16
     

2. Get your Gloo Platform Helm values file. For more information, see Upgrade guide.

   • Single cluster:

       helm get values gloo-platform -o yaml -n gloo-mesh > gloo-single.yaml
     open gloo-single.yaml
       

   • Multicluster:

       helm get values gloo-platform -n gloo-mesh -o yaml --kube-context $MGMT_CONTEXT > mgmt-plane.yaml
     open mgmt-plane.yaml
       

       helm get values gloo-platform -n gloo-mesh -o yaml --kube-context $REMOTE_CONTEXT > data-plane.yaml
     open data-plane.yaml
       

Step 4: Set up Redis authentication link

Choose how to authenticate with your Redis cluster. Depending on your Redis provider, you might be required to use a specific authentication method, such as TLS certificates or a username and password.

TLS-encryption link

For TLS-encrypted communication with Redis, the steps vary by provider.

1. Enable TLS encryption for your Redis instance.

   • For local instances, you might create your own certificates such as with openssl or cert-manager to make the public Certificate Authority (CA) certificate, ca.crt. Then, you create a secret with this CA certificate so that the Gloo client can establish trust for the TLS connection to the Redis server instance. Note that server-side credentials for a mutual TLS (mTLS) connection are currently not supported.
   • For cloud instances, the provider often manages the certificates for you. For example, AWS ElastiCache encrypts in-transit communication with their own public CA. Note: Most up-to-date apps include CA certificates that trust the AWS CA by default. However, if after setup, you notice timeouts and SSL_connect failed: certificate verify failed messages in the logs, your app might not have the CA certificates. Depending on your app, you can update the CA certificates such as by logging into the pod and running update-ca-certificates. For an example, see the Bring your own Redis deployment example in the verification guide.

2. For local instances: Create a Kubernetes secret to store the TLS credentials in the same namespace as the rate limiter. Run the following command from the local directory where your TLS certificates and key were created. Update the filenames of the certificates as necessary. Note: If you do not have certificates, such as if you use a cloud provider like AWS ElastiCache, skip this step.

     kubectl apply -f - <<EOF
   apiVersion: v1
   kind: Secret
   type: Opaque
   metadata:
     name: "redis-certs-rl"
     namespace: $GLOO_MESH_RELEASE_NAMESPACE
     labels:
       app.kubernetes.io/managed-by: Helm
     annotations:
       meta.helm.sh/release-name: $GLOO_MESH_RELEASE_NAME
       meta.helm.sh/release-namespace: $GLOO_MESH_RELEASE_NAMESPACE
   data:
     ca.crt: $(cat ca.crt | base64 | tr -d '\n')
   EOF
     

3. Prepare the redis.certs section of the Helm values file that you previously retrieved.

   For local instances, enable the redis.certs section for the rate limiter add-on. Include the details of the secret that you previously created.

     
   rateLimiter:
     redis:
       certs:
         # Set certs to true to bring your own certs.
         enabled: true
         # The 'mountPoint', 'caCert', 'signingKey', and 'secretName' are set to the default values, 
         # which you must update if you changed these in your secret.
         mountPoint: "/etc/tls"
         caCertKey: "redis.crt"
         signingKey: "redis.key"
         secretName: "redis-certs-rl"
       # Configure the Redis service to use a TLS connection with the TLS certs that you provide.
       service:
         socket: tls
     

   For cloud instances like AWS ElastiCache that use a public CA, set up the Redis to use TLS encryption. Omit the secret name and CA cert key fields, because the cloud instances use a public CA.

     
   rateLimiter:
     redis:
       certs:
         # Set certs to false because you do not bring your own certs.
         enabled: false
       # Enable using a TLS connection with the cloud Redis instance.
       service:
         socket: tls
     

Password authentication link

For a username and password auth, complete the following steps.

1. Enable password auth for your Redis instance.

2. Save the following values as environment variables.

   • Username: If your Redis instance does not have a username, use default.
   • Password: Note the password string.

     export REDIS_UN=default
   export REDIS_PW=<password>
     

3. Create a Kubernetes secret to store the credentials in the same namespace as the management server.

     kubectl apply -f- <<EOF
   apiVersion: v1
   kind: Secret
   type: Opaque
   metadata:
      name: redis-auth-secrets-rl
      namespace: $GLOO_MESH_ADDONS_NAMESPACE
      labels:
        app.kubernetes.io/managed-by: Helm
      annotations:
        meta.helm.sh/release-name: $GLOO_MESH_RELEASE_NAME
        meta.helm.sh/release-namespace: $GLOO_MESH_ADDONS_NAMESPACE
   stringData:
     #If Redis doesn't have an explicit username, specify ‘default’.
     username: "$REDIS_UN"
     password: "$REDIS_PW"
   EOF 
     

4. Prepare the redis.auth section of the Helm values file that you previously retrieved.

     
   rateLimiter:
     redis:
       auth:
         # Enable authentication for Redis with the username and password in the referenced secret.
         enabled: true
         # The name of the secret in the same namespace as the add-ons.
         secretName: redis-auth-secrets-rl
         # The key name in the data section of the Kubernetes secret
         # that has the password.
         passwordKey: password
         # The key name in the data section of the Kubernetes secret
         # that has the username.
         usernameKey: username
     

Step 5: Configure other Redis settings link

Complete the redis section of your Helm values file. For more settings, see the Helm reference.

1. To use your own Redis instance, disable the default Solo-provided Redis.

     
   rateLimiter:
     redis:
       # The default deployment of Redis for the add-on.
       # Set to false to bring your own Redis instance.
       enabled: false
     

2. Include the address to use to connect to your Redis cluster.

   • For local instances, this value is typically the Kubernetes service name that the management server can access, such as gloo-mesh-redis.gloo-mesh.
   • For a cloud instance, this value is typically the primary endpoint, such as master.my-redis-name.cache.amazonaws.com for AWS ElastiCache or 10.xxx.xx.xx for Google Cloud Memorystore.
   • You can omit the port, which defaults to 6379.

     
   rateLimiter:
     redis:
       # Provide the host to your external Redis database, such as AWS ElastiCache or Google Cloud Memorystore.
       # Note that this field does not require the port, which defaults to 6379.
   	   hostname: "master.my-redis-name.cache.amazonaws.com"
     

Step 6: Upgrade the Helm installation link

Continue with the Upgrade guide to reinstall Gloo Mesh Gateway with your own Redis instance. Include the Helm values file that you prepared in the previous steps.

The following example Helm values file shows a simple installation with the Redis section filled out for an AWS ElastiCache instance for the rate limiter.

  
rateLimiter:
  enabled: true
  redis:
    # Disable the default local Redis instance.
    enabled: false
    # Provide the host to your external Redis database, such as AWS ElastiCache or Google Cloud Memorystore.
    # Note that this field does not require the port, which defaults to 6379.
	hostname: "master.my-redis-name.cache.amazonaws.com"
    # Optionally set to true if your Redis instance runs in clustered mode.
    clustered: false
    # Optionally enable authentication for Redis with the username and password in the referenced secret.
    # The 'secretName', 'passwordKey', and 'usernameKey' are set to the default values,
    # which you must update if you changed these in your secret.
    auth:
      enabled: true
      secretName: redis-auth-secrets-rl
      passwordKey: password
      usernameKey: username
    # Setting certs to false because you don't have your own certs to use for the TLS connection.
    certs:
      enabled: false
    # Configure the Redis service to use a TLS connection with the TLS certs that you provide.
    service:
      socket: tls
  

Verify the connection link

Before you begin verifying the connection, populate the Redis instance with data. For example, you can follow the Basic rate limit policy guide to send responses. In Redis, you can use the monitor command to notice as the counters increase towards the rate limit.

To review the data that is stored in Redis, you can connect to the local Redis instance. If you bring your own Redis, you must use those connection values. For more information, see your Redis provider docs, such as Amazon ElastiCache or Google Cloud Memorystore.

Choose between using the Redis CLI or RedisInsights.

Redis CLI link

You can use the Redis CLI that is built into your deployment to review the data in your database.

1. Install Redis in your cluster so that you can use the redis-cli to connect to your Redis instance.

   When you install Gloo Mesh Gateway with the default settings, Redis is installed in the cluster for you. You can skip this step.

   To connect to your own local Redis instance, make sure to get its details, such as my-redis in a redis namespace, as well as any authentication details.

   To connect your own external Redis in a cloud instance, such as AWS ElastiCache:

   1. Deploy the Redis CLI, such as with the following simple deployment. For more information, see the Redis docs.

        kubectl apply -f - <<EOF
      apiVersion: apps/v1
      kind: Deployment
      metadata:
        name: redis-cli
        labels:
          app: redis-cli
      spec:
        replicas: 1
        selector:
          matchLabels:
            app: redis-cli
        template:
          metadata:
            labels:
              app: redis-cli
          spec:
            containers:
            - name: redis-cli
              image: redis:latest
              command: ["sleep", "infinity"]  # Keeps the container running indefinitely
      EOF
        

   2. Verify that you can access the Redis pod by logging into the pod.

        kubectl exec -it deploy/redis-cli -- /bin/sh
        

   3. Optional: For TLS connections that use a public CA, such as for AWS ElastiCache, make sure that the Redis CA certificates are up to date.

         apt-get install ca-certificates
        

   4. Verify that the CA certificates are installed.

        ls /etc/ssl/certs
        

      Example truncated output:

        002c0b4f.0
      02265526.0
      ...
      Amazon_Root_CA_1.pem
      Amazon_Root_CA_2.pem
      Amazon_Root_CA_3.pem
      Amazon_Root_CA_4.pem
      ...
        

2. From your terminal, log in to the Redis pod and open a Redis CLI shell.

   You can open a Redis CLI shell session by logging in to the Redis instance in the gloo-mesh namespace, such as gloo-mesh-redis which backs the Gloo management server and UI.

     kubectl exec -n gloo-mesh deploy/gloo-mesh-redis -it -- redis-cli
     

   You can open a Redis CLI shell session by logging in to the Redis instance in the add-ons namespace, such as redis in the gloo-mesh namespace, which backs the rate limiter, external auth, or portal server add-ons.

     kubectl exec -n gloo-mesh deploy/redis -it -- redis-cli
     

   You can open a Redis CLI shell session by logging into the Redis pod with the details that you got in the previous step. Include the --tls flag if your Redis instance enforces a TLS connection. Include the -a flag with your password if your Redis instance enforces auth.

     kubectl exec -n redis deploy/my-redis -it -- redis-cli --tls -a 'my-password'
     

   You can open a Redis CLI shell session by logging into the redis-cli pod that you deployed in the previous step. Update the host with your cloud instance details. Include the --tls flag if your Redis instance enforces a TLS connection. Include the -a flag with your password if your Redis instance enforces auth.

     kubectl exec -n default deploy/redis-cli -it -- redis-cli -h master.<my-redis>.abcdefg.use2.cache.amazonaws.com --tls -a 'my-password' -p 6379
     

3. Run Redis CLI commands, such as monitor or keys. For more information, see the Redis docs.

   Example response of data populated by the management server:

     keys *
   1) "$CLUSTER_NAME~apps~v1~Deployment"
   2) "reports#$CLUSTER_NAME#httpbin#$CLUSTER_NAME"
   3) "$CLUSTER_NAME~install.istio.io~v1alpha1~IstioOperator"
   4) "reports#$CLUSTER_NAME#gloo-mesh#globalcccfde00-e4a9-4867-b6c0-5df8601f6e1e"
   5) "$CLUSTER_NAME~admissionregistration.k8s.io~v1~MutatingWebhookConfiguration"
   6) "reports#idx"
   7) "reports#$CLUSTER_NAME#store#$CLUSTER_NAME"
   8) "gloo.mesh.key.set"
     

   Example response of rate limit counters:

     monitor
   OK
   1681223733.198361 [0 10.xxx.x.x:xxxxx] "INCRBY" "tree|solo.io|generic_key^gloo-mesh.rl-server-config-gloo-mesh-$CLUSTER_NAME-rate-limiter|generic_key^counter|1718236800" "1"
     

   Example response of API keys for the external auth service and portal server add-ons. Note that the API key name and value are hashed based on the apiKeyStorage.secretKey that you configured when setting up the add-ons.

     keys *
   1) "1bbe3f25-5a2a-4797-8ac0-d59c7480377a"
   2) "abcmHk2Iwcy3VL+NNswyrs3B4YIEx39BG84jzwcoWB8="
   
   get "1bbe3f25-5a2a-4797-8ac0-d59c7480377a"
   "xksmHk2Iwcy3VL+NNswyrs3B4YIEx39BG84jzwcoWB8="
   
   get "abcmHk2Iwcy3VL+NNswyrs3B4YIEx39BG84jzwcoWB8="
   "{\"api_key\":\"abcmHk2Iwcy3VL+NNswyrs3B4YIEx39BG84jzwcoWB8=\",\"labels\":[\"first.last@example.com\"],\"metadata\":{\"config_id\":\"gloo-mesh.api-key-auth-default-cluster-name-mgt-ext-auth-service\",\"created-ts-unix\":\"1718742070\",\"name\":\"gold-api-key-name\",\"usagePlan\":\"gold\",\"username\":\"first.last@example.com\"},\"uuid\":\"1bbe3f25-5a2a-4797-8ac0-d59c7480377a\"}"
     

RedisInsights link

RedisInsights is a desktop application that you can install for UI- and CLI-based interactions. For more information, see the Redis docs.

1. Install RedisInsights. The steps vary depending on where you want to install the app.
   • On your local machine: Download RedisInsights.
   • In your Kubernetes cluster: See the Redis docs.
2. Enable port forwarding on the Redis deployment. Update the namespace and deployment to match the Redis instance in your cluster.

     kubectl port-forward -n gloo-mesh deploy/gloo-mesh-redis 6379
     

3. Launch the RedisInsights app.
4. Click + Add Redis Database and enter in your localhost and port details, which by default are 127.0.0.1:6379.
5. Click the database that you just added.
6. Use the UI or CLI to explore the data. For example, you can click the Browser tab (the key icon) and toggle the Key view to explore all of the keys in the database instance, as shown in the following screenshot.