Vm Infrastructure Ops

Check if eth-collector systemd service is running:

gcloud compute ssh eth-realtime-collector --zone=us-east1-b \
  --command='sudo systemctl status eth-collector'

Expected Output (healthy):

● eth-collector.service - Ethereum Real-Time Collector
   Loaded: loaded (/etc/systemd/system/eth-collector.service; enabled)
   Active: active (running) since ...

Alternative (use provided script):

.claude/skills/vm-infrastructure-ops/scripts/check_vm_status.sh

2. View Logs (Live Tail)

Stream real-time logs from the collector service:

gcloud compute ssh eth-realtime-collector --zone=us-east1-b \
  --command='sudo journalctl -u eth-collector -f'

What to Look For:

• "Block inserted" messages every ~12 seconds (healthy)
• gRPC errors, DNS resolution failures (unhealthy)
• "Connection refused" or "Metadata server unreachable" (network issues)

3. View Recent Logs (Last 100 Lines)

gcloud compute ssh eth-realtime-collector --zone=us-east1-b \
  --command='sudo journalctl -u eth-collector -n 100'

4. Restart Service

Restart the collector service after configuration changes or to recover from errors:

gcloud compute ssh eth-realtime-collector --zone=us-east1-b \
  --command='sudo systemctl restart eth-collector'

Alternative (use provided script with pre-checks):

.claude/skills/vm-infrastructure-ops/scripts/restart_collector.sh

When to Use:

• After deploying code updates
• Recovering from gRPC metadata validation errors
• After Secret Manager credential updates

5. VM Hard Reset

Hard reset the VM instance (use as last resort):

gcloud compute instances reset eth-realtime-collector --zone=us-east1-b

When to Use:

• VM network connectivity completely lost
• DNS resolution failures
• Metadata server unreachable
• Service restart doesn't resolve issues

Warning: This forcefully restarts the VM. All in-memory state is lost.

6. Verify Data Flow

After restarting services, verify data is flowing to ClickHouse:

cd 
doppler run --project aws-credentials --config prd -- python3 -c "
import clickhouse_connect
import os
client = clickhouse_connect.get_client(
    host=os.environ['CLICKHOUSE_HOST'],
    port=8443,
    username='default',
    password=os.environ['CLICKHOUSE_PASSWORD'],
    secure=True
)
result = client.query('SELECT MAX(timestamp), MAX(number) FROM ethereum_mainnet.blocks FINAL')
print(f'Latest block: {result.result_rows[0][1]:,} at {result.result_rows[0][0]}')
"

Expected Output (healthy):

Latest block: 23,800,000+ at <within last 60 seconds>

Common Failure Modes

See VM Failure Modes for detailed troubleshooting guide.

Quick Reference:

Systemd Commands

See Systemd Commands Reference for complete systemd operations.

Quick Reference:

# Status
sudo systemctl status eth-collector

# Start
sudo systemctl start eth-collector

# Stop
sudo systemctl stop eth-collector

# Restart
sudo systemctl restart eth-collector

# Enable (start on boot)
sudo systemctl enable eth-collector

# Disable (don't start on boot)
sudo systemctl disable eth-collector

# View service logs
sudo journalctl -u eth-collector

# Follow logs live
sudo journalctl -u eth-collector -f

Operational History

Infrastructure Recovery (2025-11-10 07:00 UTC):

• VM network failure detected (DNS resolution failed, metadata server unreachable)
• Recovery: VM reset restored network connectivity
• eth-collector service restarted with .strip() fix (gRPC metadata validation resolved)
• Real-time data flow confirmed: blocks streaming every ~12 seconds
• Database verified: 23.8M blocks (2015-2025), latest block within seconds

Maintainability SLO Achievement: Critical infrastructure failure (VM network down) resolved in <30 minutes (VM reset + service restart + verification).

Related Documentation

• ClickHouse Migration ADR - Production database migration
• Real-Time Collector Deployment Guide - VM deployment
• Gap Monitor README - Automated gap detection
• Data Pipeline Monitoring Skill - Cloud Run Jobs monitoring

Scripts

• check_vm_status.sh - Automated status check via gcloud
• restart_collector.sh - Safe restart with pre-checks

References

• vm-failure-modes.md - Common failure scenarios and solutions
• systemd-commands.md - Complete systemd operations reference