Skip to content

Latest commit

 

History

History

advanced

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
README.md
README.md
 
 
docker-compose.png
docker-compose.png
 
 
docker-compose.yml
docker-compose.yml
 
 

README.md

Graph Node docker-compose - ADVANCED

Table of Contents

Outline

❗ This guide assumes you've completed Graph Node docker-compose - BASIC.

Now we need to improve our setup to meet "production" demands

Run separate Indexer and Query nodes

So far we have been running a graph-node in "combined-node" mode. This means it performs both indexing and serving queries.

We need to separate these functions since the demand for each may scale up/down depending on how many subgraphs are syncing, and how many requests we are serving,

Stop your existing graph-nodes

cd ~/indexer-docker-compose/graph-node/basic && docker-compose down

Navigate to this folder, update docker-compose.yml with your Postgres login and Web3 Provider info, and then start the new docker compose.

cd ~/indexer-docker-compose/graph-node/advanced
nano docker-compose.yml
docker-compose up -d

In the new docker-compose.yml we create two separate instances of graph-nodes. One in "query-node" mode, and the other in "index-node" mode.

services:
  graph-node-query:
    # ...
    environment:
      node_role: "query-node"
      node_id: "query-node"
  graph-node-indexer:
    # ...
    ports:
      - "8100:8000" # http
      - "8120:8020" # json-rpc
      - "8140:8040" # metrics
    environment:
      node_role: "index-node"
      node_id: "index-node"
      BLOCK_INGESTOR: "index-node"

Things to pay attention to:

  • Deploying subgraphs is now handled by the "index-node", while queries are handled by the "query-node".
  • We are binding the "index-node" to a different set of ports eg. 81XX. Your create/deploy commands will need to reflect this port change.
  • In the BASIC guide we set Postgres to store data in ~/subgraph-data/postgres. Since we do the same here, we won't lose any existing subgraph sync data (unless you changed your authentication)

Add Websockets & Health Monitoring

Update Nginx config

In order to support websockets and health monitoring, we must change our Nginx config. Replace the indexer.conf you created in the BASIC guide to the final one in the /nginx folder of this repo. You will need to update the server_name.

sudo cp ~/indexer-docker-compose/nginx/indexer.conf /etc/nginx/sites-enabled

nano /etc/nginx/sites-enabled/indexer.conf
# Update server_name

Next update /etc/nginx/nginx.conf to add support for Nginx connection upgrades as follows:

sudo sudo nano /etc/nginx/nginx.conf
http {
  # Add this code block within "http"
  map $http_upgrade $connection_upgrade {
          default upgrade;
          ''      close;
  }
}

Excellent! Now we have support for websockets, and we can perform health checks against the query node. Let's put the changes into effect:

sudo nginx -t
sudo systemctl reload nginx

We also need to generate SSL certs to acces https://indexer.mysite.com/index-node/ from the browser

Now we need Certbot to issue a certificate.

# Install
sudo apt-get install software-properties-common
sudo add-apt-repository ppa:certbot/certbot
sudo apt-get update
sudo apt-get install python-certbot-nginx

# Run
sudo certbot --nginx certonly

# If successful, restart Nginx
sudo systemctl reload nginx

Things to pay attention to:

  • Health checks are performed against index-node/
  • This config only exposes the "query-node" on ports 80XX. If you want external access to your "index-node", you will need to add that.

Check Indexer health

To check the health and status of a particular subgraph, we use the index-node/graphql endpoint. Here is an example of how to check the subgraph jannis/gravity:

Property value
URL http://indexer.mysite.com/index-node/graphql
Request type POST 
# Query body
{
  indexingStatusForCurrentVersion(subgraphName: "jannis/gravity") {
    synced
    health
    fatalError {
      message
      block {
        number
        hash
      }
      handler
    }
    chains {
      chainHeadBlock {
        number
      }
      latestBlock {
        number
      }
    }
  }
}

You should get a response like this:

{
  "data": {
    "indexingStatusForCurrentVersion": {
      "chains": [
        {
          "chainHeadBlock": {
            "number": "10637299"
          },
          "latestBlock": {
            "number": "10637299"
          }
        }
      ],
      "fatalError": null,
      "health": "healthy",
      "synced": true
    }
  }
}

💯 PRO-TIP: Learn more about health checks here

Harden the Postgres Database

The Graph team does not recommend running postgres using docker-compose, since it needs to be very stable. I have not needed to do this yet, however I will list your options just in case:

  • Run Postgres as systemd service on your server
  • Use a 3rd-party provider

Bonus: Deploy your own subgraph

In this example we will use the subgraph jannis/gravity to demonstrate how you would deploy your own subgraph to your indexer.

First install graph-cli, download the desired subgraph repo, and make any changes to the subgraph.yaml

# Install the graph-cli
npm i -g @graphprotocol/graph-cli

# Use any existing subgraph
git clone https://github.com/Jannis/gravity-subgraph.git && cd gravity-subgraph

Now we are ready to push our subgraph to our indexer.

In package.json, add the following scripts (don't forget to add a comma in the line above)

"create-indexer": "graph create jannis/gravity --node http://127.0.0.1:8020",
"deploy-indexer": "graph deploy jannis/gravity --debug --ipfs https://testnet.thegraph.com/ipfs/ --node http://127.0.0.1:8020"

Now generate the files, and deploy

yarn
yarn codegen
yarn create-indexer
yarn deploy-indexer

If successful, you will see Deployed to http://127.0.0.1:8000/subgraphs/name/jannis/gravity/graphql. Check that your subgraph is syncing using docker logs, as mentioned above, and happy querying!

Extras

Useful Commands

You may find these useful to check how you server is performing

# Check Memory
free -m
ps -o pid,user,%mem,command ax | sort -b -k3 -r

# Check Storage
ncdu

# Hardware monitoring dashboard in a terminal
# https://github.com/bcicen/ctop
docker run --rm -ti \
  --name=ctop \
  --volume /var/run/docker.sock:/var/run/docker.sock:ro \
  quay.io/vektorlab/ctop:latest

Next Steps

You are absolutely crushing it! 💪

Continue to Monitoring Infra