DOAS – Database Observation and Alert System

Lightweight monitoring tool that inspects database clusters and sends e-mail alert if required. This tool was developed to proactively detect the most common issues of database clusters in Kubernetes and enable faster response by system administrators.

doas.sh is executed on schedule to check database engines defined in values.yaml for:

• Pod health and restart status
• Memory usage, CPU Usage and Throttle
• PVC disk usage

If any threshold is exceeded or an issue is detected, an e-mail is sent.

Setup

1. Build Docker image and push to a registry.
2. Update values.yaml and smtpsecrets.yaml files.
3. Deploy the helm chart.

helm install doas . --namespace doas --create-namespace --values values.yaml

Helm chart

├── .gitignore
├── Chart.yaml
├── values.yaml
├── docker/
│   ├── .dockerignore
│   └── Dockerfile
├── scripts/
│   └── doas.sh
└── templates/
    ├── clusterrole.yaml
    ├── clusterrolebinding.yaml
    ├── configmap.yaml
    ├── deployment.yaml
    ├── smtpsecrets.yaml
    ├── serviceaccount.yaml
    └── _helpers.tpl

Notes

• Deployment is used instead of CronJob to avoid creating a new pod for every execution.
• This tool was primarily designed for monitoring database cluster health, but applicable to any application deployed on Kubernetes.