Product Information
Removal of Prometheus/Grafana-based monitoring in SAP BTP, Kyma runtime
I’m part of the team working on the Observability feature for SAP BTP, Kyma runtime, and in my previous blog post: Deprecation of Prometheus/Grafana based monitoring in SAP BTP, Kyma runtime, I announced the deprecation of the monitoring functionality based on Prometheus and Grafana. Today, I’d like to inform you about the imminent removal of the functionality with Kyma release 2.20 (planned for November 2023).
Consequences
The removal of the integrated Prometheus/Grafana installation affects the following features:
- You will lose access to Grafana and the ability to browse metrics for up to 30 days.
- You will lose access to the Prometheus API and the ability to access metrics for up to 30 days.
Still, the following options will not be affected:
- You can manually install the classic in-cluster monitoring solution. Read Installing a custom Prometheus stack in Kyma for more details.
- You can continue bringing your custom solution for the collection and long-term storage of metrics inside or outside the cluster.
- You can continue integrating with commercial monitoring offerings as for example outlined in the mission Extend a Kyma Based Multitenant Application With Day 2 Operations.
Conclusion
Kyma’s feature for storing and browsing metrics based on Prometheus and Grafana will be removed from the SAP BTP, Kyma runtime in version 2.20.
If you haven’t integrated your solution with an external monitoring system yet, this is the right time to start evaluating and implementing it: For example, integrate with a 3rd-party provider like Dynatrace, or implement a custom installation of Prometheus/Grafana.
Stay tuned for any upcoming SAP BTP, Kyma runtime blog posts and follow the related topic page.
Hi Andreas,
Is the log collection within the Kyma system still handled by Fluent Bit? Or is this now an OTel Collector as well?
The diagrams in the initial announcement indicated that Fluent Bit was in place but this would be moved to an OTel collector eventually. I'm trying to understand the current state.
I am comparatively new to OpenTelemetry within Kubernetes/Kyma, and I want to validate that the attached diagram is largely a valid future state for starting down a path today.
A napkin-sketch future state diagram for observability within Kyma that I am trying to ensure is possible
Thanks!