We will take a look at the available options for logging in ROSA. As ROSA does not come preconfigured with a logging solution, we can easily set one up. In this section review the install proceedure for the EFK (Elasticsearch, Fluentd and Kibana) stack (via Operators), then we will take a look at three methods with which one can view their logs.
- We will look at the logs directly through the pod using
- We will forward the logs to AWS CloudWatch and view them from there.
- We will use Kibana (EFK Stack) to search our logs.
The cluster logging components are based upon Fluentd, (and Elasticsearch and Kibana, if deployed). The collector, Fluentd, is deployed to each node in the cluster. It collects application logs and writes them to Elasticsearch (ES) or forwards it to CloudWatch. Kibana is the centralized, web UI where users and administrators can create rich visualizations and dashboards with the aggregated data. We will also look at using AWS CloudWatch as well.
Installing the Cluster Logging Addon service#
In these steps we will install the logging addon service to forward our logs; in our case to CloudWatch. One can install the service though the OCM UI or by using the CLI. We will use the OCM UI below. If you want to use the CLI you can see the steps here.
- Access the OCM UI, select your cluster, and click on the Add-ons tab.
Click on the Cluster Logging Operator
Select the logs you want to collect. If you want to forward somewhere other than CloudWatch leave that box unchecked. You can select the defaults and leave the region blank (unless you want to use a different region). Click Install.
It will take about 10 minutes to install.
Output data to the streams/logs#
- Output a message to stdout Click on the Home menu item and then click in the message box for "Log Message (stdout)" and write any message you want to output to the stdout stream. You can try "All is well!". Then click "Send Message".
- Output a message to stderr Click in the message box for "Log Message (stderr)" and write any message you want to output to the stderr stream. You can try "Oh no! Error!". Then click "Send Message".
View application logs using
- Go to the CLI and enter the following command to retrieve the name of your frontend pod which we will use to view the pod logs:
$ oc get pods -o name pod/ostoy-frontend-679cb85695-5cn7x pod/ostoy-microservice-86b4c6f559-p594d
So the pod name in this case is ostoy-frontend-679cb85695-5cn7x.
oc logs ostoy-frontend-679cb85695-5cn7xand you should see your messages:
$ oc logs ostoy-frontend-679cb85695-5cn7x [...] ostoy-frontend-679cb85695-5cn7x: server starting on port 8080 Redirecting to /home stdout: All is well! stderr: Oh no! Error!
You should see both the stdout and stderr messages.
View logs with CloudWatch#
- Access the web console for your AWS account and go to CloudWatch.
Click on Logs > Log groups in the left menu to see the different groups of logs depending on what you selected during installation. If you followed the previous steps you should see 2 groups. One for
<cluster-name>-XXXXX-applicationand one for
Click on the log stream for the "frontend" pod. It will be titled something like
Filter for "stdout" and "stderr" the expand the row to show the message we had entered earlier along with much other information.
We can also see other messages in our logs from the app. Enter "microservice" in the search bar, and expand one of the entries. This shows us the color recieved from the microservice and which pod sent that color to our frontend pod.
You can also use some of the other features of CloudWatch to obtain useful information. But how to use CloudWatch is beyond the scope of this tutorial.
View logs with Kibana#
NOTE: In order to use EFK, this section assumes that you have successfully completed the steps outlined in Installing OpenShift Logging.
Run the following command to get the route for the Kibana console:
oc get route -n openshift-logging
Open up a new browser tab and paste the URL. You will first have to define index patterns. Please see the Defining Kibana index patterns section of the documentation for further instructions on doing so.
Familiarization with the data#
In the main part of the console you should see three entries. These will contain what we saw in the above section (viewing through the pods). You will see the stdout and stderr messages that we inputted earlier (though you may not see it right away as we might have to filter for it). In addition to the log output you will see information about each entry. You can see things like:
- namespace name
- pod name
- host ip address
- log level
You will also see that there is data from multiple sources and multiple messages. If we expand one of the twisty-ties we can see further details
Let's look for any errors encountered in our app. Since we have many log entries (most from the previous networking section) we may need to filter to make it easier to find the errors. To find the error message we outputted to stderr lets create a filter.
- Click on "Add a filter+" under the search bar on the upper left.
- For "Fields..." select (or type) "level"
- For "Operators" select "is"
- In "Value..." type in "err"
- Click "Save"
You should see now only one row is returned that contains our error message.
NOTE: If nothing is returned, depending on how much time has elapsed since you've outputted the messages to the stdout and stderr streams you may need to set the proper time frame for the filter. If you are following this lab consistently then the default should be fine. Otherwise, in the Kibana console click on the top right where it should say "Last 15 minutes" and click on "Quick" then "Last 1 hour" (though adjust to your situation as needed).