TheGraph - Public Cloud Abstraction

Automated deployment using Terraform, Kubernetes and Helm

Video Guide: https://www.youtube.com/watch?v=ORr3yWBSn90

Introduction

This is a Graph Protocol automated deployment using the following tools:

• Terraform - used to bring up the necessary cloud infrastructure
• Kubernetes - well known platform for managing containerized workloads and services
• Helm - a Kubernetes deployment tool for automating creation, configuration, and deployment of applications to Kubernetes clusters

Public Clouds Supported:

Public Cloud	Supported
Azure Cloud	YES
AWS	YES
GKE	YES

The aim of the project is to provide a fully automated method of deploying and managing TheGraph services while keeping the much needed modularity that the protocol needs.

This guide is also a work in progress and any feedback aiming to improve will be appreciated.

HOW TO USE IT

There are three simple steps (curently only Azure is supported):

1. Install the tools (prerequisites)
2. Deploy the Cloud Infrastructure using Terraform
3. Deploy TheGraph containers using Helm

If you are a first time user we are strongly recommending to read through this guide and try to understand the differences step by step.

---

1. Install prerequisites

Tool	Recommended version
Terraform	0.14.x - latest patch (now: 0.14.11)
Kubectl	Latest (now: v1.21.1)
Helm	Latest (now: v3.6.0)

1.A. The short version (commands used, in order):

cd $HOME
apt update && apt install curl wget unzip git

# Terraform
wget https://releases.hashicorp.com/terraform/0.14.11/terraform_0.14.11_linux_amd64.zip
unzip terraform_0.14.11_linux_amd64.zip
chmod 755 terraform
mv terraform /usr/local/bin/terraform

# Kubectl
curl -LO "https://dl.k8s.io/release/$(curl -L -s https://dl.k8s.io/release/stable.txt)/bin/linux/amd64/kubectl"
chmod 755 kubectl
mv kubectl /usr/local/bin/kubectl

# Helm
wget https://get.helm.sh/helm-v3.6.0-linux-amd64.tar.gz
tar -xzvf helm-v3.6.0-linux-amd64.tar.gz
chmod 755 linux-amd64/helm
mv linux-amd64/helm /usr/local/bin/helm

1.B. The long version (with command output and comments):

In order to install the following prerequisites make sure that you are logged as root.

cd $HOME
apt update && apt install curl wget unzip git

Install Terraform

wget https://releases.hashicorp.com/terraform/0.14.11/terraform_0.14.11_linux_amd64.zip
unzip terraform_0.14.11_linux_amd64.zip
chmod 755 terraform
mv terraform /usr/local/bin/terraform

# Confirm it's working:
terraform version
Terraform v0.14.11

Install kubectl

curl -LO "https://dl.k8s.io/release/$(curl -L -s https://dl.k8s.io/release/stable.txt)/bin/linux/amd64/kubectl"
chmod 755 kubectl
mv kubectl /usr/local/bin/kubectl

# Confirm it's working:
kubectl version
Client Version: version.Info{Major:"1", Minor:"21", GitVersion:"v1.21.1", GitCommit:"5e58841cce77d4bc13713ad2b91fa0d961e69192", GitTreeState:"clean", BuildDate:"2021-05-12T14:18:45Z", GoVersion:"go1.16.4", Compiler:"gc", Platform:"linux/amd64"}

Install Helm

wget https://get.helm.sh/helm-v3.6.0-linux-amd64.tar.gz
tar -xzvf helm-v3.6.0-linux-amd64.tar.gz
chmod 755 linux-amd64/helm
mv linux-amd64/helm /usr/local/bin/helm

# Confirm it's working:
helm version
version.BuildInfo{Version:"v3.6.0", GitCommit:"7f2df6467771a75f5646b7f12afb408590ed1755", GitTreeState:"clean", GoVersion:"go1.16.3"}

---

2. Deploy your cloud infra using terraform

You can find the deployment instructions, depending on you chosen cloud provider, here:

• DEPLOY ON AZURE
• DEPLOY ON AWS
• DEPLOY ON Google Cloud

---

3. Deploy TheGraph containers using helm

After deploying with terraform, the following steps are the same, independend of your chosen cloud provider.

3.A. The short version (commands used, in order):

# Install
cd graphprotocol
helm repo add grafana https://grafana.github.io/helm-charts
helm repo add prometheus-community https://prometheus-community.github.io/helm-charts
helm dependency build .

# Edit the graphprotocol/values.yaml and install your deployment
helm install thegraph-deployment .

# Check
helm list
kubectl get pods

# Get the IPs
kubectl get ingress

3.B. The long version (with command output and comments):

# Go the helm charts directory and install the dependencies
cd graphprotocol
helm repo add grafana https://grafana.github.io/helm-charts
helm repo add prometheus-community https://prometheus-community.github.io/helm-charts
helm dependency build .

>>> Edit the graphprotocol\values.yaml file. <<< The values.yaml file contains all the variables needed by TheGraph in order to work. Follow the comments inside the file in order to identify the variables that needs to be changed.

# Deploy TheGrap apps (remember to edit the values.yaml file before)
helm install thegraph-deployment .

# Confirm that the status of the deployment is "deployed"
helm list
NAME                    NAMESPACE       REVISION        UPDATED                                 STATUS          CHART                           APP VERSION
ingress                 default         1               2021-06-16 15:35:18.666415289 +0000 UTC deployed        ingress-nginx-3.33.0            0.47.0
thegraph-deployment     default         1               2021-06-16 17:33:00.757097124 +0000 UTC deployed        graphprotocol-agent-0.1.0       0.1.0


# Check the status of the pods while the pods are beying created (imediately after issueing the helm deploy command)
kubectl get pods
NAME                                                           READY   STATUS              RESTARTS   AGE
index-node-0                                                   0/1     ContainerCreating   0          59s
indexer-agent-6b8787-z2gc7                                     0/1     ContainerCreating   0          59s
indexer-service-7887975b5c-fmhvs                               0/1     ContainerCreating   0          59s
ingress-ingress-nginx-controller-8f9b6b667-qf2z6               1/1     Running             0          118m
query-node-85457498fd-59tlz                                    0/2     ContainerCreating   0          59s
query-node-85457498fd-k5kz8                                    0/2     ContainerCreating   0          59s
thegraph-deployment-grafana-7d85cb59bc-m4zlz                   0/2     Init:0/1            0          59s
thegraph-deployment-kube-state-metrics-97bdc4d9c-ccv6q         0/1     ContainerCreating   0          59s
thegraph-deployment-prometheus-alertmanager-687dcf5b7f-ffg6k   0/2     ContainerCreating   0          59s
thegraph-deployment-prometheus-node-exporter-g2zxz             1/1     Running             0          59s
thegraph-deployment-prometheus-pushgateway-856dc495c-sjpgn     0/1     ContainerCreating   0          59s
thegraph-deployment-prometheus-server-585594ddd-x262r          0/2     ContainerCreating   0          59s

NOTE:

Sometimes the indexer-service gets up faster than the index-node and it will fail on the first attempt. It will automatically reload by itself when the index-node will be ready. Due to this behaviour, in the next output we can notice a "3" on the RESTARTS column corespondig to the indexer-service.

# The initial deployment may take up to 5 minutes, when it's done all the pods should be running:
kubectl get pods
NAME                                                           READY   STATUS    RESTARTS   AGE
index-node-0                                                   1/1     Running   0          52m
indexer-agent-6b8787-jm8hq                                     1/1     Running   0          40m
indexer-service-7887975b5c-fmhvs                               1/1     Running   3          52m
ingress-ingress-nginx-controller-8f9b6b667-qf2z6               1/1     Running   0          170m
query-node-85457498fd-59tlz                                    2/2     Running   0          52m
query-node-85457498fd-k5kz8                                    2/2     Running   0          52m
thegraph-deployment-grafana-7d85cb59bc-m4zlz                   2/2     Running   0          52m
thegraph-deployment-kube-state-metrics-97bdc4d9c-ccv6q         1/1     Running   0          52m
thegraph-deployment-prometheus-alertmanager-687dcf5b7f-ffg6k   2/2     Running   0          52m
thegraph-deployment-prometheus-node-exporter-g2zxz             1/1     Running   0          52m
thegraph-deployment-prometheus-pushgateway-856dc495c-sjpgn     1/1     Running   0          52m
thegraph-deployment-prometheus-server-585594ddd-x262r          2/2     Running   0          52m

# Get the public IP of the ingress controller and add them to your DNS Zone and/or your hosts file
kubectl get ingress
NAME                         CLASS    HOSTS                                    ADDRESS         PORTS   AGE
indexer-agent                <none>   agent.thegraphtest.easy2stake.com        20.93.201.127   80      50m <- only if ingress is enabled. Disabled by default
indexer-service              <none>   service.thegraphtest.easy2stake.com      20.93.201.127   80      50m
new-test-grafana             <none>   grafana.thegraphtest.easy2stake.com      20.93.201.127   80      50m
new-test-prometheus-server   <none>   prometheus.thegraphtest.easy2stake.com   20.93.201.127   80      50m <- only if ingress is enabled. Disabled by default.

Access the ingress endpoints in your browser and confirm it's working.

Credentials for Grafana:

Grafana user: admin Grafana password: kubectl get secret --namespace default thegraph-deployment-grafana -o jsonpath="{.data.admin-password}" | base64 --decode ; echo

Access Prometheus & Indexer Agent

Prometheus: No user and password defined. Use the whitelist inside values.yaml to control the access in case you activated the ingress for it, or access prometheus using PORT Forwarding: kubectl get pods kubectl port-forward pods/prometheus-server-<pod_id> 9090:9090

Browse: http://127.0.0.1:9090/

Indexer Agent access: Use the whitelist inside values.yaml to control the access in case you activated the ingress for it, or access Indexer Agent using PORT Forwarding: kubectl get pods kubectl port-forward pods/indexer-agent-<pod_id> 8000:8000

Browse: http://127.0.0.1:8000/

Security:

We strongly recommend that you enable HTTPS for all your services published via INGRESS controller. Once you have DNS records created for mentioned hosts (see above), you can use Cert-Manager (https://cert-manager.io/docs/) along with Let's Encrypt Certficate Authority (https://letsencrypt.org)

4. Troubleshooting

Work in progress.