[BUG] agent crashing

[kevinburkesegment]

Agent Environment

7.50.3-jmx, running on amd64 in Kubernetes.

Describe what happened:

Agents crash with the following stack trace (every goroutine gets dumped, but this is the most germane). I can't tell if this is an issue with the Go runtime or with the Datadog agent doing something it's not supposed to with unsafe.

I ran the Datadog agent tests with the race detector and gcflags=-d=checkptr turned on, and found one data race, but it didn't match this one.

log="goroutine 1 [chan receive, 905 minutes]:" pod=datadog-agent-4b6c6 time=2024-12-14T02:18:24.601865825Z
log= pod=datadog-agent-4b6c6 time=2024-12-14T02:18:24.601857714Z
log=" /usr/local/go/src/runtime/mgc.go:178 +0x6b" pod=datadog-agent-4b6c6 time=2024-12-14T02:18:24.601855714Z
log="created by runtime.gcenable" pod=datadog-agent-4b6c6 time=2024-12-14T02:18:24.601853354Z
log=" /usr/local/go/src/runtime/asm_amd64.s:1598 +0x1 fp=0xc0000ad7e8 sp=0xc0000ad7e0 pc=0x22f5d41" pod=datadog-agent-4b6c6 time=2024-12-14T02:18:24.601850844Z
log=runtime.goexit() pod=datadog-agent-4b6c6 time=2024-12-14T02:18:24.601848654Z
log=" /usr/local/go/src/runtime/mgc.go:178 +0x26 fp=0xc0000ad7e0 sp=0xc0000ad7c8 pc=0x229d946" pod=datadog-agent-4b6c6 time=2024-12-14T02:18:24.601846334Z
log=runtime.gcenable.func1() pod=datadog-agent-4b6c6 time=2024-12-14T02:18:24.601843643Z
log=" /usr/local/go/src/runtime/mgcsweep.go:299 +0xff fp=0xc0000ad7c8 sp=0xc0000ad780 pc=0x22a889f" pod=datadog-agent-4b6c6 time=2024-12-14T02:18:24.601828623Z
log=runtime.bgsweep(0x0?) pod=datadog-agent-4b6c6 time=2024-12-14T02:18:24.601826162Z
log=" /usr/local/go/src/runtime/mgcsweep.go:391 +0xf9 fp=0xc0000ad780 sp=0xc0000ad730 pc=0x22a8ad9" pod=datadog-agent-4b6c6 time=2024-12-14T02:18:24.601823882Z
log=runtime.sweepone() pod=datadog-agent-4b6c6 time=2024-12-14T02:18:24.601821272Z
log=" /usr/local/go/src/runtime/mgcsweep.go:629 +0x34d fp=0xc0000ad730 sp=0xc0000ad640 pc=0x22a910d" pod=datadog-agent-4b6c6 time=2024-12-14T02:18:24.601818552Z
log="runtime.(*sweepLocked).sweep(0xc2a1460?, 0x0)" pod=datadog-agent-4b6c6 time=2024-12-14T02:18:24.601806681Z
log=" /usr/local/go/src/runtime/mgcsweep.go:846 +0x2e5 fp=0xc0000ad640 sp=0xc0000ad5c0 pc=0x22a9e45" pod=datadog-agent-4b6c6 time=2024-12-14T02:18:24.601804461Z
log=runtime.(*mspan).reportZombies(0x7f367858daa8) pod=datadog-agent-4b6c6 time=2024-12-14T02:18:24.601802051Z
log=" /usr/local/go/src/runtime/panic.go:1047 +0x5d fp=0xc0000ad5c0 sp=0xc0000ad590 pc=0x22bd07d" pod=datadog-agent-4b6c6 time=2024-12-14T02:18:24.601799781Z
log="runtime.throw({0x7c84b92?, 0xc000000400?})" pod=datadog-agent-4b6c6 time=2024-12-14T02:18:24.601796871Z
log="goroutine 3 [running]:" pod=datadog-agent-4b6c6 time=2024-12-14T02:18:24.601793230Z
log= pod=datadog-agent-4b6c6 time=2024-12-14T02:18:24.601777690Z
log="fatal error: found pointer to free object" pod=datadog-agent-4b6c6 time=2024-12-14T02:18:24.599670983Z
log="0xc000000000 free marked zombie" pod=datadog-agent-4b6c6 time=2024-12-14T02:18:24.599371515Z
log="runtime: marked free object in span 0x7f367858daa8, elemsize=1024 freeindex=0 (bad use of unsafe.Pointer? try -d=checkptr)" pod=datadog-agent-4b6c6 time=2024-12-14T02:18:24.599312012Z

Describe what you expected:

I expected the agent to continue running or shut down in an orderly fashion.

Steps to reproduce the issue:

I don't know how to reproduce it.

Additional environment details (Operating System, Cloud provider, etc):