Send new GitHub mention to Slack as message
Slack + GitHub
Send new GitHub mention to Slack as message
Slack + GitHub
If you are running into issues with your NGINX Ingress controller, it could be due to one of several potential causes. In this guide, we'll walk you through the troubleshooting steps to identify your issue.
If you are troubleshooting an issue with your Ingress and you know it isn’t an infrastructure issue, then you likely have an issue with your Ingress controller.
In this guide, we’re going to outline how you can identify the specific issue impacting your NGINX Ingress controller.
Here are some of the common causes for why an NGINX Ingress controller may not be working properly:
To debug your particular scenario, there are several steps to gathering useful information. These next steps assume that your Ingress controller is deployed in the namespace nginx-ingress and <nginx-ingress-pod> is the name of one of the pods.
When you run into issues, checking the logs is often the first step to getting more information. By doing this, you’ll be able to see if the Ingress controller is failing to start, which could mean you have misconfigured RBAC or a missing default server TLS Secret.
To check the logs for both the Ingress controller software and the NGINX access and error logs, you can run this command:
You can manage the degree of detail you receive back about the Ingress controller by using the -v command-line argument for ingress controller software logs (-v=4: most detailed, -v=1: least detailed). To make a similar verbosity adjustment for the NGINX logs, you need to do that by configuring corresponding ConfigMap keys.
If the logs show that the Ingress controller is starting correctly, then you can move on to the next likely cause: invalid values of annotations.
To check if this is your issue, you should review the controller logs, then check the events of your Ingress resource to see whether your NGINX configuration was successfully applied.
You can check the events of an Ingress resource by running the following command:
If you see a Normal event with the reason AddedOrUpdated, it means that the configuration was successfully applied.
The Ingress controller creates a main NGINX configuration file for each Ingress/VirtualServer resource and stores it in the /etc/nginx/conf.d folder. You can also verify that your annotation values are correct by checking the configuration file, which you can print by running the following command:
If you have verified that your annotations have the correctly listed values; then you should next check if you have an invalid VirtualServer or VirtualServerRoute.
You can next check if your configuration was successfully applied by NGINX to your VirtualServer or VirtualServiceRoute resources. To check if this is your issue, you can utilize the controller logs, the generated configuration file, and check the events for the VirtualServer and VirtualServerRoute resources.
You can check events for a VirtualServer resource with this command:
You can check events for a VirtualServerRoute resource with this command:
Similar to checking the events of an Ingress Resource, you are looking for Normal events with the reason AddedOrUpdated to signal that the configuration was successfully applied.
Now that you have checked the logs, the generated configuration file, and the events of the VirtualServers, it may be an issue of an invalid Policy.
You can check the events of a Policy resource using the following command:
If you see Normal events with the reason AddedOrUpdated, it means that the Policy resource was successfully accepted by the Ingress controller. You will need to check the VirtualServer events that reference that policy to know if the policy was actually applied.
So far, we’ve narrowed down the list of potential issues by verifying which steps are operating as expected. If the configuration has not been applied, and you have checked the other causes, then it is likely an issue with the values for the ConfigMap keys.
To check this, you can use the logs, the generated configuration file, and also the events of the ConfigMap.
You can check the events of the ConfigMap using the following command:
With this, and the other methods described above, you should be able to gather all the information you need to determine why the configuration is not being applied correctly.
What if your Ingress controller is starting successfully and configuration is applied correctly, but you are getting unexpected resources back? You might have unhealthy backend pods or an issue with your backend service.
You can check for backend pod or service issues by reviewing the logs, the generated configuration file, checking the live activity dashboard, and running NGINX in debug mode.
We’ve already covered the commands for the first two methods, so let’s move on to the live activity dashboard.
The dashboard is enabled and available on port 8080. As long as your nginx-status command-line argument is not set to false, you’ll be able to run the following command to forward connections to port 8080 on your local machine to port 8080 of your NGINX Plus Ingress Controller pod:
You can also read about more customization options for the dashboard.
While it’s rare that the issue is due to a bug in NGINX code, you can check for issues by using the debug logs. You can enable them with these two steps:
1. Set error-log-level to debug in ConfigMap.
2. Use -nginx-debug command-line argument when running the ingress controller.
Troubleshooting an NGINX configuration issue requires checking for multiple potential causes and running various commands to get more information.
There’s an easier and faster way to manage your Kubernetes troubleshooting.
With Blink, you can automate this process and run all these steps with a simple click. You can also customize and expose your automation as a self-service app that your team members can use. Instead of fielding a support ticket, you can enable other developers to solve their own NGINX problem, guided by your automation.
Create your free Blink account today and streamline your troubleshooting today.
Transform your security and platform operations today with 5000+ no-code automations.