Kubernetes Observability for SREs with Dynatrace

The cloud-native community provides a constant stream of amazing new projects enabling new ways to leverage k8s for our current and future workloads. When it comes to observability, SREs and DevOps can leverage:
- Prometheus to collect metrics
- Fluentbit to collect logs
- Event exporter to get the events from Kubernetes
- Jaeger and OpenTelemetry to collect distributed traces
Dynatrace is keeping pace with the observability and automation demands of DevOps and SRE by leveraging the latest k8s APIs to ingest and analyze all pillars of modern observability: logs, metrics, traces, events, and more …
In this Performance Clinic Henrik Rexed, Cloud Native Advocate at Dynatrace, explains how Dynatrace:
- Collects natively metrics, logs, and traces from your cluster
- Ingests metrics from open-source products like Fluentbit, OpenTelemetry, and Prometheus
- Enables you to manage your SLO based on metrics coming from k8s events or logs

[@shajunair5310]

Superb

[@BhavinShah74]

Hi.. I can see on ur demo 51:20 services are linked to distributed traces.. would you like to share what needs to do for same..?

[@isitobservable]

Hi, The traces showed at 51:20 is generated by KSPAN. KSPAN generates Opentelemetry traces based on Kubernetes Event. Unfortunately, I forgot to stop my Chaos experiments during the webinar...as a consequence Istio and my demo apps were suffering. So I was not able to show any traces generated by ISTIO. But if you enable the tracing in Istio, you should be able to ingest the generated traces. Here is the Github repo used for this Performance Clinic : github.com/dynatrace-perfclinics/K8s_observability_4_SRE

[@wyattblake4126]

I thought that Dynatrace gave you everything so you don’t need to ingest Prometheus. Maybe you talk about this in the video (I haven’t finished it yet)

[@isitobservable]

Hi, Dynatrace will automatically collects logs, events , metrics from your cluster. It will collect metrics related to the application observed by Dynatrace. Lots of technologies are exposing metrics in a Prometheus Format ( backup solution, service mesh, CICD systems...and more) . So ingesting external metrics allow you to extend your observability .Let me know if you need more information.