Dynatrace Managed Disk Overuse and Missing Kubernetes Service Monitoring

[HuseynNajafli]

Hello everyone,
I would like to share two issues I’ve been facing and I hope you can suggest solutions for them.

Issue 1: Disk problem in Dynatrace Managed

I frequently encounter disk usage spikes on my / partition. Up to 5 GB of space is being filled within a single day. When I check the logs, they don’t seem to consume much space. The main space-consuming component appears to be the repeated downloads of OneAgent versions. I do not want these versions to download automatically. I have already disabled Auto-Update and set it to manual, yet the versions are still being downloaded every day. Moreover, older versions are being downloaded as well, even though none of them are being used. Every day when I come to work, I face the same issue and manually Exclude those old versions from the UI.

How can I prevent these versions from being automatically downloaded? Also, in general, what are the primary components that consume the most disk space in Dynatrace Managed (specifically on the / partition)?

---

Issue 2: Kubernetes services and other services not being monitored

As we know, after installing OneAgent on Kubernetes, I need to restart my microservices (applications or simply “services”) to enable monitoring. After the initial OneAgent installation, I did perform the restart and all my services appeared in the monitoring dashboard.

However, a strange issue occurs:
After updating Ubuntu OS on my Dynatrace and Kubernetes nodes, I restarted those nodes — which is expected. But afterwards, I noticed that my Kubernetes services were no longer visible. To restore monitoring, I again had to restart all microservices in Kubernetes.

Additionally, due to the disk issue mentioned above, one of my Dynatrace nodes crashed. After expanding the / partition and restarting the node, everything started working normally again — except that Kubernetes services disappeared once more. This again required restarting the microservices.

How can I fix this problem? How can I ensure that my services remain monitored continuously? Restarting a node should not force me to restart hundreds of microservices just to restore monitoring.

Below are the screenshots showing:

• When my services are visibleBefore 

• When my services disappearAfter 

I look forward to your suggestions. Thank you!

[jason_gs]

hi
Point #1 - raise a ticket for Dynatrace Support and ask for the auto download feature to be disabled. This has to be done from the Dynatrace side.  

[Mizső]

 Hi @HuseynNajafli ,

What is the instrumentation type of on your Kubernetes? calssicFullStack or cloudNativeFullStack. CalssicFullStack has a race condition. At cluster restarts (during upgrade) we also experienced this behavior at calssicFullStack.

https://docs.dynatrace.com/docs/shortlink/how-it-works-k8s-classic-fullstack#classic-limitations

I hope it helps.

Best regards,

János