Performing Timeline Reconstruction With Plaso

Performing Timeline Reconstruction with Plaso

When to Use

• When building a comprehensive forensic timeline from multiple evidence sources
• For correlating events across file system metadata, event logs, browser history, and registry
• During complex investigations requiring chronological reconstruction of activities
• When standard log analysis is insufficient to establish the sequence of events
• For presenting investigation findings in a visual, chronological format

Detection Gaps & Validation

A super-timeline's greatest weakness is that it makes noise and forged timestamps look identical to ground truth. Guard against these:

• Timezone and clock errors corrupt the whole timeline. Plaso stores events in UTC, but parsers infer source timezones; run log2timeline.py with the correct --timezone and confirm the image's TimeZoneInformation registry value. A wrong offset or a host whose clock was manually changed silently shifts events, fabricating false sequences. State the timezone basis in every report.
• Super-timeline noise buries the signal. A full-image run emits millions of events; $MFT/$UsnJrnl alone can dominate. Without a --filter-file or psort date-range/source filter you will drown. Pivot from known anchors (malware execution, login) and expand outward rather than reading top-to-bottom.
• MACB timestamps are stompable and parser-dependent. Anti-forensic tooling backdates $STANDARD_INFORMATION times that Plaso surfaces; the harder-to-forge $FILE_NAME ($FN) times come from a separate MFT parse. Compare $SI vs $FN to spot timestomping, and corroborate a file's "creation" against USN Journal, Prefetch/Amcache execution, and event logs before trusting it.
• Coverage gaps = missing parsers/artifacts, not absence of activity. If you ran a targeted --parsers list, anything outside it is invisible; deleted/rotated logs and VSS snapshots won't appear unless included. "No evidence in the timeline" must be qualified by which parsers and sources were actually processed.
• Interpretation false positives. A single artifact's timestamp ≠ user action: $MFT access times are unreliable on modern Windows, antivirus/indexing/backup touch files en masse, and Prefetch reflects execution by any caller. Confirm a pivotal event with at least two independent artifact types (e.g., Prefetch + 4688 + LNK) before placing it in the narrative.

Prerequisites

• Plaso (log2timeline/psort) installed on forensic workstation
• Forensic disk image(s) in raw (dd), E01, or VMDK format
• Sufficient storage for Plaso output (can be 10x+ the image size)
• Minimum 8GB RAM (16GB+ recommended for large images)
• Timeline Explorer (Eric Zimmerman) or Timesketch for visualization
• Understanding of timestamp types (MACB: Modified, Accessed, Changed, Born)

Workflow

Step 1: Install Plaso and Prepare the Environment

# Install Plaso on Ubuntu/Debian
sudo add-apt-repository ppa:gift/stable
sudo apt-get update
sudo apt-get install plaso-tools

# Or install via pip
pip install plaso

# Or use Docker (recommended for dependency isolation)
docker pull log2timeline/plaso

# Verify installation
log2timeline.py --version
psort.py --version

# Create output directory
mkdir -p /cases/case-2024-001/timeline/

# Verify the forensic image
img_stat /cases/case-2024-001/images/evidence.dd

Step 2: Generate the Plaso Storage File with log2timeline

# Basic processing of a disk image (all parsers)
log2timeline.py \
   --storage-file /cases/case-2024-001/timeline/evidence.plaso \
   /cases/case-2024-001/images/evidence.dd

# Process with specific parsers for faster targeted analysis
log2timeline.py \
   --parsers "winevtx,prefetch,mft,usnjrnl,lnk,recycle_bin,chrome_history,firefox_history,winreg" \
   --storage-file /cases/case-2024-001/timeline/evidence.plaso \
   /cases/case-2024-001/images/evidence.dd

# Process with a filter file to focus on specific paths
cat << 'EOF' > /cases/case-2024-001/timeline/filter.txt
/Windows/System32/winevt/Logs
/Windows/Prefetch
/Users/*/NTUSER.DAT
/Users/*/AppData/Local/Google/Chrome
/Users/*/AppData/Roaming/Mozilla/Firefox
/$MFT
/$UsnJrnl:$J
/Windows/System32/config
EOF

log2timeline.py \
   --filter-file /cases/case-2024-001/timeline/filter.txt \
   --storage-file /cases/case-2024-001/timeline/evidence.plaso \
   /cases/case-2024-001/images/evidence.dd

# Using Docker
docker run --rm -v /cases:/cases log2timeline/plaso log2timeline \
   --storage-file /cases/case-2024-001/timeline/evidence.plaso \
   /cases/case-2024-001/images/evidence.dd

# Process multiple evidence sources into one timeline
log2timeline.py \
   --storage-file /cases/case-2024-001/timeline/combined.plaso \
   /cases/case-2024-001/images/workstation.dd

log2timeline.py \
   --storage-file /cases/case-2024-001/timeline/combined.plaso \
   /cases/case-2024-001/images/server.dd

Step 3: Filter and Export Timeline with psort

# Export full timeline to CSV (super-timeline format)
psort.py \
   -o l2tcsv \
   -w /cases/case-2024-001/timeline/full_timeline.csv \
   /cases/case-2024-001/timeline/evidence.plaso

# Export with date range filter (focus on incident window)
psort.py \
   -o l2tcsv \
   -w /cases/case-2024-001/timeline/incident_window.csv \
   /cases/case-2024-001/timeline/evidence.plaso \
   "date > '2024-01-15 00:00:00' AND date < '2024-01-20 23:59:59'"

# Export in JSON Lines format (for ingestion into SIEM/Timesketch)
psort.py \
   -o json_line \
   -w /cases/case-2024-001/timeline/timeline.jsonl \
   /cases/case-2024-001/timeline/evidence.plaso

# Export with specific source type filters
psort.py \
   -o l2tcsv \
   -w /cases/case-2024-001/timeline/registry_events.csv \
   /cases/case-2024-001/timeline/evidence.plaso \
   "source_short == 'REG'"

psort.py \
   -o l2tcsv \
   -w /cases/case-2024-001/timeline/evtx_events.csv \
   /cases/case-2024-001/timeline/evidence.plaso \
   "source_short == 'EVT'"

# Export for Timeline Explorer (dynamic CSV)
psort.py \
   -o dynamic \
   -w /cases/case-2024-001/timeline/timeline_explorer.csv \
   /cases/case-2024-001/timeline/evidence.plaso

Step 4: Analyze Timeline with Timesketch

# Install Timesketch (Docker deployment)
git clone https://github.com/google/timesketch.git
cd timesketch
docker compose up -d

# Import Plaso file into Timesketch via CLI
timesketch_importer \
   --host http://localhost:5000 \
   --username analyst \
   --password password \
   --sketch_id 1 \
   --timeline_name "Case 2024-001 Workstation" \
   /cases/case-2024-001/timeline/evidence.plaso

# Alternatively, import JSONL
timesketch_importer \
   --host http://localhost:5000 \
   --username analyst \
   --sketch_id 1 \
   --timeline_name "Case 2024-001" \
   /cases/case-2024-001/timeline/timeline.jsonl

# In Timesketch web UI:
# 1. Search for events: "data_type:windows:evtx:record AND event_identifier:4624"
# 2. Apply Sigma analyzers for automated detection
# 3. Star/tag important events
# 4. Create stories documenting the investigation narrative
# 5. Share with team members

Step 5: Perform Targeted Timeline Analysis

# Analyze specific time periods around known events
python3 << 'PYEOF'
import csv
from collections import defaultdict
from datetime import datetime

# Load incident window timeline
events_by_hour = defaultdict(list)
source_counts = defaultdict(int)

with open('/cases/case-2024-001/timeline/incident_window.csv', 'r', errors='ignore') as f:
    reader = csv.DictReader(f)
    total = 0
    for row in reader:
        total += 1
        timestamp = row.get('datetime', row.get('date', ''))
        source = row.get('source_short', row.get('source', 'Unknown'))
        description = row.get('message', row.get('desc', ''))

        source_counts[source] += 1

        # Group by hour for activity patterns
        try:
            dt = datetime.strptime(timestamp[:19], '%Y-%m-%dT%H:%M:%S')
            hour_key = dt.strftime('%Y-%m-%d %H:00')
            events_by_hour[hour_key].append({
                'time': timestamp,
                'source': source,
                'description': description[:200]
            })
        except (ValueError, TypeError):
            pass

print(f"Total events in incident window: {total}\n")

print("=== EVENTS BY SOURCE TYPE ===")
for source, count in sorted(source_counts.items(), key=lambda x: x[1], reverse=True):
    print(f"  {source}: {count}")

print("\n=== ACTIVITY BY HOUR ===")
for hour in sorted(events_by_hour.keys()):
    count = len(events_by_hour[hour])
    bar = '#' * min(count // 10, 50)
    print(f"  {hour}: {count:>6} events {bar}")

# Find hours with unusual activity spikes
avg = total / max(len(events_by_hour), 1)
print(f"\n=== ANOMALOUS HOURS (>{avg*3:.0f} events) ===")
for hour in sorted(events_by_hour.keys()):
    if len(events_by_hour[hour]) > avg * 3:
        print(f"  {hour}: {len(events_by_hour[hour])} events (SPIKE)")
PYEOF

Key Concepts

Tools & Systems

Common Scenarios

Scenario 1: Ransomware Attack Reconstruction Process the full disk image with Plaso, filter to the week before encryption was discovered, identify the initial access vector from browser history and event logs, trace privilege escalation through registry and Prefetch, map lateral movement from network logon events, pinpoint encryption start from MFT timestamps showing mass file modifications.

Scenario 2: Data Theft Investigation Create super-timeline from suspect's workstation, filter for USB device connection events, file access timestamps, and cloud storage browser activity, build a narrative showing data staging, compression, and exfiltration, present timeline to legal team with tagged evidence points.

Scenario 3: Multi-System Breach Analysis Process disk images from all affected systems into a single Plaso storage file, import into Timesketch for collaborative analysis, search for lateral movement patterns across system timelines, identify the patient-zero system and initial compromise vector, map the full attack chain across the environment.

Scenario 4: Insider Threat After-Hours Activity Filter timeline to non-business hours only, identify file access patterns outside normal working times, correlate with authentication events (badge access, VPN logon), search for data access to sensitive directories during these periods, build evidence package for HR/legal.

Output Format

Timeline Reconstruction Summary:
  Evidence Sources:
    Disk Image: evidence.dd (500 GB, NTFS)
    Plaso Storage: evidence.plaso (2.3 GB)

  Processing Statistics:
    Total events extracted: 4,567,890
    Parsers used: 45 (winevtx, prefetch, mft, usnjrnl, lnk, chrome, firefox, winreg, ...)
    Processing time: 3h 45m

  Incident Window (2024-01-15 to 2024-01-20):
    Events in window: 234,567
    Event Sources:
      MFT:          89,234
      Event Logs:   45,678
      USN Journal:  56,789
      Registry:     23,456
      Prefetch:     1,234
      Browser:      5,678
      LNK Files:    2,345
      Other:        10,153

  Key Timeline Events:
    2024-01-15 14:32 - Phishing email opened (browser)
    2024-01-15 14:33 - Malicious document downloaded
    2024-01-15 14:35 - PowerShell executed (Prefetch + Event Log)
    2024-01-15 14:36 - C2 connection established (Registry + Event Log)
    2024-01-16 02:30 - Mimikatz execution (Prefetch)
    2024-01-16 02:45 - Lateral movement to DC (Event Log)
    2024-01-17 03:00 - Data exfiltration (MFT + USN Journal)
    2024-01-18 03:00 - Log clearing (Event Log)

  Exported Files:
    Full Timeline:     /timeline/full_timeline.csv (4.5M rows)
    Incident Window:   /timeline/incident_window.csv (234K rows)
    Timesketch Import: /timeline/timeline.jsonl