Cert Alert!

This repo contains a Docker image and kubernetes deployment to monitor the expiration date of SSL certificates. Expiration dates are logged to stdout as well as DataDog. Cert Alert might be a bit of a misnomer since it's assumed that the user will set up alerting on the DataDog side. There is currently no built-in alerting.

Installation

Step 1: Create a ConfigMap with your hosts

The list of hosts to check is controlled by a ConfigMap, which you'll need to create first with the following command:

kubectl create configmap cert-alert-config --from-file=hosts.txt

Your hosts.txt file should be a flat list of hostnames. For example:

beta.popitup.com
popitup.com
scaryphotobooth.com
stopandpayus.com
www.popitup.com
www.scaryphotobooth.com
www.stopandpayus.com

Step 2: Create the deployment

Once your ConfigMap is created you can create the deployment:

kubectl apply -f deployment.yaml

And that's it! Logging will start immediately.

Updating host configuration

If you need to update the list of hosts in the future you can just delete and re-create the ConfigMap and deployment:

kubectl delete deployment cert-alert-deployment
kubectl delete configmap cert-alert-config
kubectl create configmap cert-alert-config --from-file=hosts.txt
kubectl apply -f deployment.yaml

DataDog integration

By default, cert_alert.py tries to send metrics to DataDog. The metric is named ssl_expiries.days_to_expiry and is tagged with the hostname being checked. This metric can be used to set up a monitor so that you be alerted when you have a certificate approaching its expiration date.

Currently it's assumed that DataDog is running as a DaemonSet in your cluster. That way, the endpoint can be auto-discovered by querying the AWS instance metadata API. If that doesn't work, it will use localhost.

PRs welcome if this approach doesn't work in your environment :).

zeus911 / cert-alert