Just like GKE (and unlike EKS), AKS comes with an integrated logging solution. All we have to do is enable one of the AKS addons. To be more precise, we'll enable the monitoring addon. As the name indicates, the addon does not fulfill only the needs to collect logs, but it also handles metrics. However, we are interested just in logs. I believe that nothing beats Prometheus for metrics, especially since it integrates with HorizontalPodAutoscaler. Still, you should explore AKS metrics as well and reach your own conclusion. For now, we'll explore only the logging part of the addon.
1 az aks enable-addons 2 -a monitoring 3 -n devops25-cluster 4 -g devops25-group
The output is a rather big JSON with all the information about the newly enabled monitoring addon. There's nothing exciting in it.
It's important to note that we could have enabled the addon when we created the cluster by adding -a monitoring argument to the az aks create command.
If you're curious what we got, we can list the Deployments in the kube-system Namespace.
1 kubectl -n kube-system get deployments
The output is as follows.
NAME DESIRED CURRENT UP-TO-DATE AVAILABLE AGE heapster 1 1 1 1 1m kube-dns-v20 2 2 2 2 1h kubernetes-dashboard 1 1 1 1 1h metrics-server 1 1 1 1 1h omsagent-rs 1 1 1 1 1m tiller-deploy 1 1 1 1 59m tunnelfront 1 1 1 1 1h
The new addition is the omsagent-rs Deployment that will ship the logs (and metrics) to Azure Log Analytics. If you describe it, you'll see that it is based on microsoft/oms image. That makes it the first and the only time we switched from Fluentd to a different log shipping solution. We'll use it simply because Azure recommends it.
Next, we need to wait for a few minutes until the logs are propagated to Log Analytics. This is the perfect moment for you to take a short break. Go fetch a cup of coffee.
Let's open Azure portal and see Log Analytics in action.
1 open "https://portal.azure.com"
Please click the All services item from the left-hand menu, type log analytics in the Filter field, and click the Log Analytics item.
Unless you are already using Log Analytics, there should be only one active workspace. If that's the case, click it. Otherwise, if there are multiple workspaces, choose the one with the ID that matches the id entry of the az aks enable-addons output.
Click the menu item Logs in the General section.
Next, we'll try to limit the output entries only to those that contain random-logger. Please type the query that follows in the Type your query here... field.
1 ContainerLog | where Name contains "random-logger"
Click the Run button, and you'll be presented with all the random-logger entries.
By default, all the fields are shown in the table, and many of them are either not used, or not very useful. The extra columns probably distract us from absorbing the logs, so we'll change the output.
It's easier to specify which columns we need, than which ones we don't. Please expand the Columns list, and click the SELECT NONE button. Next, select LogEntry, Name, and TimeGenerated fields and, once you're finished, contract the Columns list.
What you see in front of you are logs limited to random-logger and presented only through the three columns we selected.
I'll let you explore Log Analytics features on your own. Even though Azure portal's UI is not as intuitive as it could be, I'm sure you'll manage to get your way around it. If you choose to adopt AKS integration with Log Analytics, you should probably explore Log Analytics query language (https://docs.microsoft.com/en-us/azure/azure-monitor/log-query/query-language) documentation that will help you write more complex queries than the one we used.
Given that there is at least one more solution we should explore before we choose the one that fits your needs the best, we'll disable the addon. Later on, if you do like Log Analytics more than the alternatives, all you'll have to do is to enable it again.
1 az aks disable-addons 2 -a monitoring 3 -n devops25-cluster 4 -g devops25-group