[BUG] journald log start_position=forceEnd doesn't take effect

[at-wat]

Agent Environment

At least on datadog-agent 7.46.0 and 7.48.1
Tested with systemd 246.15 and 253.7

Describe what happened:
When journald log start_position is set to forceEnd, all old logs are uploaded on agent start.

conf.d/journald.yaml

logs:
  - type: journald
    start_position: forceEnd
    include_units:
      ...

Describe what you expected:
Only logs after starting datadog-agent is uploaded when setting start_position: forceEnd

Steps to reproduce the issue:

1. Set journald log start_position config to forceEnd
2. Start datadog-agent
3. Check uploaded logs (logs before agent start time is uploaded)

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

Identified bug details
Journald tailer calls journal.SeekHead and journal.SeekTail,

datadog-agent/pkg/logs/tailers/journald/tailer.go

Lines 200 to 205 in 73b8b37

	if mode == config.ForceBeginning {
	return t.journal.SeekHead()
	}
	if mode == config.ForceEnd {
	return t.journal.SeekTail()
	}

but document of sdjournal says:
https://pkg.go.dev/github.com/coreos/go-systemd/[email protected]/sdjournal#Journal.SeekHead

This call must be followed by a call to Next before any call to Get* will return data about the first element.

https://pkg.go.dev/github.com/coreos/go-systemd/[email protected]/sdjournal#Journal.SeekTail

This call must be followed by a call to Previous before any call to Get* will return data about the last element.

I confirmed forceEnd behavior is fixed by adding journal.Previous after journal.SeekTail.
seqsense/datadog-agent-alpine#106

I'll open a PR to fix this.