dotnet-monitor memory dumps not removed from diag folder

[Snehaannadurai2110]

Description

We are running dotnet-monitor alongside a .NET application in Kubernetes.

We have configured dump file collection such that the dumps are first stored in a temporary location (/diag folder) and then uploaded to an Azure Storage account.

However, the memory dump files collected by dotnet-monitor are not being cleaned up from the /diag folder after being uploaded.

Over time, these files accumulate and consume significant disk space, causing:

• Node disk pressure
• Pod evictions
• Potential application instability

---

Observed Disk Usage

Below are examples of disk usage inside different pods:

kubectl exec -it dotnet-app-XXXXX -- du -sh /diag
9.1G    /diag

kubectl exec -it dotnet-app-XXXXX -- du -sh /diag
32G     /diag

kubectl exec -it dotnet-app-XXXXX -- du -sh /diag
30G     /diag

---

Impact

• Dump files continuously accumulate
• No automatic cleanup observed
• Node disk space gets exhausted
• Kubernetes evicts pods due to disk pressure

---

Expected Behavior

Dump files should either:

• Be automatically cleaned up after collection, or
• Support a configurable retention/size limit to prevent disk exhaustion.

---