Installing UP9 on Minikube or Kubernetes

Installing UP9 on Minikube or Kubernetes requires three steps:

  1. Sign up & Install
  2. Select namespaces or services
  3. Restart the selected namespaces or services

Step (1/3): Sign Up & Install

To install, please sign up using this link: https://up9.app/signup. Once logged in, the system will present a ready-to-run Helm command.

UP9 installation by Helm Please make sure you choose the correct Helm version, copy this text, and run it in your terminal window.

Step (2/3): Select Namespaces or Services

A few seconds after your pods are running, your browser screen should change, asking you to select the namespaces or services that UP9 will observe, listening to their traffic.

Selecting namespaces or services Go ahead and select the namespaces or services and press “Apply.”

Step (3/3): Restart Your Containers

For the UP9 sidecars to install themselves into the target pods, a rolling restart must be performed for the target namespaces or services. Please restart the namespaces or services by running a command similar to this one:

Copy to clipboard
kubectl rollout restart deployments -n <target-namespace>

Support

For support, feel free to use any one of the three:

  • Schedule time with an UP9 staff member using this link.
  • Email support@up9.com
  • Use our in-app messaging application to start a conversation.

Troubleshooting

Questions & Answers

I followed the installation instructions and I don’t see the service selection screen

Check that the UP9 agent is showing heartbeats

In the UP9 dashboard, find “Clusters” under “Organization Settings,” and make sure you see both the agent and the collectors with a recent 'Last Seen' date.

Kubernetes Clusters

If you don't see the clusters, or you don't see any indication of heartbits, follow the next instruction.

Check that the UP9 containers are functioning

Run the kubectl command and see if there are two UP9 pods, and make sure that they are running.

Kubectl

Other Reasosns

If the containers seem to work, but there are no signs of life in the dashboard, it probably means there’s a connectivity issue between the cluster and UP9 (e.g. no firewall rule set). If you cannot seem to figure out what’s preventing connectivity, or if the UP9 containers are failing for any reason, please contact support.

UP9 seems to be installed correctly, but it looks like I'm stuck in the “Waiting for Traffic” state

Once UP9 is installed, it expects traffic to flow in the system. UP9 works by observing service traffic. The more traffic UP9 observes, the more information it has to train its models. If there isn't enough traffic, UP9 can not generate a model. To generate traffic, you can simply use the system or run pre-existing end-to-end tests.

UP9 seems to be installed correctly, but it looks like I'm stuck in the “Rollout” state

After you've installed UP9,in order for the UP9 sidecars to install themselves into the target pods, a rolling restart needs to be performed for the target namespaces or services. Please restart the namespaces or services by running a command similar to this one:

Copy to clipboard
kubectl rollout restart deployments -n <target-namespace>

To make sure changes have gone into effect, each of the target services should have 2/2 services in their pods, where the second one is the UP9 sidecar.

sidecars

UP9 seems to be installed correctly, but it looks like I'm stuck in the “Generating” Model state

If everything else seems to work, it could be that not enough traffic is flowing through the system. To generate traffic, you can simply use the system or run pre-existing end-to-end tests.