Leverage monitoring in Kyma runtime to better operate and troubleshoot your business extensions

With release 1.17, we have enabled access to the monitoring capability in Kyma via the Grafana UI. This will provide additional support to developers to troubleshoot any issues with their business extensions or event triggers.

NOTE: At the moment, Grafana provides Kyma runtime users only with read-only access. This means it is not possible to define and save custom dashboards or provide custom metrics from the applications.

To access the Grafana UI, navigate to Diagnostics –> Metrics. It uses the same authentication as the Kyma Console UI.

Kyma comes bundled with a lot of out-of-the-box dashboards. You can start accessing them right away to monitor and troubleshoot your extensions, microservices, functions, and Event triggers.

The best way to start is by exploring the available dashboards.

Let’s look at some of the available dashboards and see what metrics they offer.

Functions

In the Kyma / Functions dashboard, you can get a holistic view of how a function is behaving. These are some of the available basic metrics:

• Request duration or latency
• Resource usage
• Error rate

What about my microservcies?

The best way to start is to check out the Istio / Service dashboard to access all the deployed microservices.

Simply select the microservice from the dropdown list.

Standard metrics are provided out of the box to the developer thanks to the Istio service mesh.

If you have set up a deployment with multiple replicas, you can have a look at the Kubernetes / Compute Resources / Workload dashboard to see how various replicas are behaving resource-wise. Select namespace and deployment you want to monitor.

Further on, you can drill down and inspect each and every Pod of a given ReplicaSet to get more details.

Kubernetes metrics

There is a whole bunch of Kubernetes metrics available for you so that you could constantly keep an eye on your cluster.

In general, the Kyma runtime is managed and operated by SAP. Still, the curious folks can explore various dashboards on their own to see how their clusters behave.

One out of many available dashboards is Kubernetes / Compute Resources / Cluster.

Keep an eye on those events

To troubleshoot any event delivery- or latency-related issues, Kyma runtime provides a whole set of useful dashboards.

The most interesting one for developers is Kyma / Event Mesh / Broker-Trigger. It contains details about how events are delivered to your microservice or function. You can select a specific Namespace in which the event trigger has been configured.

To dig in further, you can refer to the Kyma / Event Mesh / Latency dashboard for any latency issues.

When you want to confirm if the events are coming from a particular connected system or not, you can check out the Kyma / Event Mesh / Delivery dashboard.

Let’s see some logs

Loki (the Kyma logging component) is also integrated into Grafana. Thanks to it, you can access the logs for your microservice or function directly from the Grafana UI.

It also provides nice search capabilities with autocompletion for various fields such as Namespaces, labels, container/function names, and so on.

Run your own queries

In case the existing dashboards and panels don’t meet your needs and you want to do some aggregations or calculations that are not available, you can easily do so by exploring Grafana and providing your own PromQL queries.

That’s not all!

That was just a quick overview, but there are many more dashboards shipped with Kyma runtime out of the box. They can help you to observe, analyze, and act on various Kyma components and Kubernetes resources. Enjoy!