Maker Monitor

This project aims to make useful Maker and Dai system metrics easily available through open source monitoring tools. I have chosen to use Prometheus as the backend for simplicity, and Grafana for the frontend for flexibility.

Developing

This project uses Docker, Kubernetes, and Tilt to try and get local development as close to the deployment as possible. Please make sure you have these tools installed before moving forward. Once you're setup we can start by creating the Kubernetes resources we only need once

# Create a namespace to operate in
k create namespace maker-monitor

# Set the namesapce in the current context to apply to all subsequent commands
k config set-context --current --namespace=maker-monitor

# This is secrect, and isn't shared in the repo, which is why you need to create it yourself
k create configmap monitor-config --from-literal=INFURA_PROJECT_ID=<your infura project ID>

# This stores the prometheus config into a k8s config map
k create configmap prometheus-config --from-file=prometheus/prometheus.yml

# Create our local PVs and PVCs to store persistant data
k create -f prometheus/k8s/vol_local.yml
k create -f grafana/k8s/vol_local.yml

Of course, this assumes you've aliased kubectl as k. Next start up the stack using Tilt!

tilt up

Tilt should open a browser UI so you can see the status of services. When all of the services have finished initialization you should be able to view Prometheus at https://localhost:9090

Building the monitor process binary

./monitor/scripts/build.sh

Note: This creates a Docker images tagged ddaws/maker-monitor:latest

Exposing Grafana via a local Ingress

To expose Grafana locally via an Ingress you'll need to make sure you have an Ingress Controller running on your local cluster. You can use the Nginx Ingress Controller which is tested and working on Mac. Please follow the instructions found here to install the local Nginx Ingress Controller. It may take a few minutes for the Ingress controller to start.

To Do

Deployments

• Convert naked pods to deployments to allow for better management (should allow for updating pods via scaling down & up)
• Mount Grafana config via a ConfigMap to version control it here in the repo
• Create remote config map for Grafana config in prod
• Add remote volume for Grafana

Metrics Collection

• Update monitor process to use the Multicall contract and query Maker stats per ETH block published

Improvements

• Update Docker image to build more efficiently (I think it's go get'ing on build)
• Maybe update Tiltfile dev image adding live syncing (requires install tools into image)

Deployment

To deploy Maker Monitor you'll need to create a series of config maps, pods and services via Kuberenetes. Let's get started by creating our local resources. You only need to do this once

# This is secrect, and isn't shared in the repo, which is why you need to create it yourself
k create configmap monitor-config --from-literal=INFURA_PROJECT_ID=<your infura project ID>

# Deploy the monitor service
k apply -f monitor/k8s/service.yml

# Deploy the prometheus service
k create configmap prometheus-config --from-file=prometheus/prometheus.yml
k apply -f prometheus/k8s/service.yml

Of course, this assumes you've aliased kubectl as k. If not, I recommend it ;)

If you didn't run into any errors you should be able to connect to Prometheus via

k port-forward prometheus-pod 9090:9090

Useful Links

• Go Ethereum Book
• Go Prometheus Client Lib
• Annotated Contracts via Hypothesis