[BUG] "Failed to find _dd_crashtracker_receiver" being outputted everywhere, including to grub bootloader

[omarkb93]

Agent Environment

Agent 7.70.2 - Commit: 83bbd72 - Serialization version: v5.0.161 - Go version: go1.24.6

Describe what happened:
My datadog installations on a couple VMs have gotten screwed up somehow. The datadog installation is automated with ansible. On a batch of VMs I created, something seems to have gone very wrong and datadog seems to be struggling with something called "dd_crashtracker" which I don't see referenced anywhere. Something is outputting the following messages to stdout at various points:

Failed to find _dd_crashtracker_receiver
Failed to start crashtracker: failed to construct crashtracker configuration
Failed: command not found

For example, this happens on ssh login

It also happens when I run conda, and a couple other odd commands on the command line. It happens when sudo apt upgrade tries to install landscape-common which seems to cause it to fail.

Most amusingly / scarily, it appears to have dumped it into a place where the grub bootloader read and then tried to make linux images based on it.

I don't really know what's going on. Any context or advice for debugging would be helpful.

Describe what you expected:

Not this

Steps to reproduce the issue:

???

Additional environment details (Operating System, Cloud provider, etc):
Ubuntu 24.04 running on Proxmox VE

[omarkb93]

I can't repeat the issue; it occurred on all 4 vms I made in a batch, but I just tested remaking one and it no longer suffers from this issue. The original behavior seems worrysome enough to try to track down though, so I still have the old vms around for debugging.