Tyk Performance Testing

This project is still a WIP. Schema might change and suggestions are welcome.

How to use

Locally

The Repo can be cloned locally. You can stand up aks, eks, or gke using the respective folders or bring in your own cluster.

Clusters

Terraform managed clusters

Once the clusters are created, you will need to add the connection config to your .kube/config. Here are examples per k8s services

AKS

az aks get-credentials \
   --resource-group "pt-westus" \
   --name "pt-westus"

EKS

aws eks --region "us-west-1" update-kubeconfig --name "pt-us-west-1"

GKE

gcloud container clusters get-credentials pt-us-west1-a \
   --zone us-west1-a \
   --project ce-team-zaid

Feature matrix

Gateway	Database Analytics	Prometheus Analytics	AuthToken	JWT	Quota	Rate Limiting	Open Telemetry
Tyk	✅	✅	✅	✅	✅	✅	✅
Kong	❌	✅	✅	❌	Implemented through Rate Limiting	✅	✅
Gravitee	✅	✅	✅	❌	Implemented through Rate Limiting	✅	❌

Self-managed cluster requirements

To run the tests on your own cluster, you will need node labels that can house the different deployments and resources.

Example:

kubectl label nodes minikube node=dependencies
kubectl label nodes minikube-m02 node=tyk
kubectl label nodes minikube-m02 node=tyk-upstream
kubectl label nodes minikube-m03 node=tyk-tests
kubectl label nodes minikube-m04 node=tyk-resources
kubectl label nodes minikube-m05 node=kong
kubectl label nodes minikube-m06 node=kong-upstream
kubectl label nodes minikube-m07 node=kong-tests
kubectl label nodes minikube-m08 node=kong-resources
kubectl label nodes minikube-m09 node=gravitee
kubectl label nodes minikube-m10 node=gravitee-upstream
kubectl label nodes minikube-m11 node=gravitee-tests
kubectl label nodes minikube-m12 node=gravitee-resources

Note: dependencies are required, while the rest are required based on where the tests are being run.

Deployments

Once you connect to a k8s cluster. You can run the deployment modules to set up the testing environments. Here is an example for the options available:

kubernetes_config_context = "performance-testing"

analytics_database_enabled    = false
analytics_prometheus_enabled  = false
auth_enabled                  = false
auth_type                     = "authToken"
quota_enabled                 = false
quota_rate                    = 999999
quota_per                     = 3600
rate_limit_enabled            = false
rate_limit_rate               = 999999
rate_limit_per                = 60
open_telemetry_enabled        = false
open_telemetry_sampling_ratio = "0.5"

hpa_enabled                 = false
hpa_max_replica_count       = 10
replica_count               = 1
hpa_avg_cpu_util_percentage = 80
external_traffic_policy     = "Local"
resources_requests_cpu      = "0"
resources_requests_memory   = "0"
resources_limits_cpu        = "0"
resources_limits_memory     = "0"

tyk_enabled          = true
tyk_version          = "v5.6"
tyk_license          = ""
tyk_deployment_type  = "Deployment"
tyk_service_type     = "ClusterIP"
tyk_go_gc            = 1600
tyk_go_max_procs     = 8
tyk_profiler_enabled = false

kong_enabled         = false
kong_version         = "3.8"
kong_deployment_type = "Deployment"
kong_service_type    = "ClusterIP"

gravitee_enabled         = false
gravitee_version         = "4.5"
gravitee_deployment_type = "Deployment"
gravitee_service_type    = "ClusterIP"

grafana_service_type = "ClusterIP"

Tests

Once the environment is set up you can run the test's module. Here is an example of the options available:

kubernetes_config_context = "performance-testing"

tyk_enabled      = true
kong_enabled     = false
gravitee_enabled = false

tests_fortio_options = "size=20"
tests_executor       = "constant-arrival-rate"
tests_duration       = 15
tests_rate           = 20000
tests_virtual_users  = 50
tests_parallelism    = 1