Funded by the Wave 1 grant, Chainstack team has built a solution to deploy customizable Graph nodes in the public clouds (which can be extended to any Kubernetes cluster), as well as monitor and operate the deployment with Kubernetes and Grafana dashboards.
Supported deployment modes:
- Network mode — the node participates in the network, processes external requests and gets the commission.
- Standalone mode — the node doesn’t join in The Graph network, processes only your own subgraph and requests.
Tools used:
- Terraform
- Helm and Helmfile
- Kubernetes Dashboard
- Prometheus and Grafana
Supported cloud providers:
- Amazon Web Services (AWS)
- Google Cloud Platform (GCP)
- Microsoft Azure
Our solution highlights:
- Our indexer Helm charts support configuration files for The Graph node. This means that you can use database replicas and configure multiple providers for blockchain protocols.
- All the components except the blockchain nodes are in Kubernetes. This means that our automation can be used in any public cloud or a self-hosted Kubernetes cluster that provides persistent volumes and load balancers.
- The standalone mode comes with its own IPFS instance and allows you to host your subgraphs without participating in the network. For those cases, when you just need to have your own indexer for your project.
GitHub repository: GitHub - chainstack/graph-deployment: The Graph related stuff to run it in the cloud