fix(cgroup): filter name=* subsystems on cgroupv1

[aymericDD]

What does this PR do?

• Fixes a bug

On cgroupv1, /proc/PID/cgroup contains a name=systemd named hierarchy. Using this key as a mount directory name produces an invalid path (/sys/fs/cgroup/name=systemd/...) that doesn't exist on disk, causing EnterPid to fail before any resource controller is joined.

Filter name=* entries at construction time in newAllCGroupManager so only real resource controllers reach EnterPid.

The regression was introduced in #704, which created cgroup/manager_linux.go and the initial newAllCGroupManager implementation that reads all subsystems from /proc/PID/cgroup without filtering named hierarchy entries.

The following disruptions were impacted on cgroupv1 systems and are now fixed:

• CPU Stress: injection completely failed — Join(pid) → EnterPid() would error on the invalid name=systemd path
• Disk Pressure: Write("blkio", ...) throttle settings could fail
• Network Disruption: Write("net_cls", ...) classid assignment could fail
• CPU Pressure: ReadCPUSet() could be affected

Code Quality Checklist

• The documentation is up to date.
• My code is sufficiently commented and passes continuous integration checks.
• I have signed my commit (see Contributing Docs).

Testing

• I leveraged continuous integration testing
  • by depending on existing unit tests or end-to-end tests.
  • by adding new unit tests or end-to-end tests.
• I manually tested the following steps:
  • Verified EnterPid no longer fails on cgroupv1 systems with name=systemd entries.
  • locally.
  • as a canary deployment to a cluster.

[datadog-datadog-prod-us1-2]

✅ Tests

🎉 All green!

❄️ No new flaky tests detected
🧪 All tests passed

🎯 Code Coverage (details)
• Patch Coverage: 100.00%
• Overall Coverage: 39.29% (+0.00%)

_{This comment will be updated automatically if new data arrives.
🔗 Commit SHA: 1355b86 | Docs | Datadog PR Page | Was this helpful? Give us feedback!}