vm-infrastructure-ops

What problem does it solve?

This Skill provides a structured approach to diagnosing and recovering Ethereum real-time data collection on a GCP e2-micro VM. It targets VM service outages, systemd failures, streaming interruptions, and network problems, enabling quick restoration of data flow.

Core Features & Use Cases

• Service health checks: verify eth-collector daemon is active and responsive.
• Live log visibility: stream or view the collector logs to spot errors.
• Recovery actions: restart the service, pull fresh credentials, or reinitialize networking.
• VM recovery options: perform a hard VM reset if networking is completely broken.
• Data flow verification: validate that blocks are flowing into ClickHouse after recoveries.

Quick Start

Use the cloud-based commands and scripts to diagnose and recover:

• Check service status on the VM: gcloud compute ssh eth-realtime-collector --zone=us-east1-b --command='sudo systemctl status eth-collector'
• View live logs: gcloud compute ssh eth-realtime-collector --zone=us-east1-b --command='sudo journalctl -u eth-collector -f'
• Restart the collector service: gcloud compute ssh eth-realtime-collector --zone=us-east1-b --command='sudo systemctl restart eth-collector'
• Quick status using provided script: .claude/skills/vm-infrastructure-ops/scripts/check_vm_status.sh
• Alternative restart script with pre-checks: .claude/skills/vm-infrastructure-ops/scripts/restart_collector.sh
• VM hard reset (last-resort): gcloud compute instances reset eth-realtime-collector --zone=us-east1-b
• Verify data flow to ClickHouse after restart: doppler run --project aws-credentials --config prd -- python3 -c "import clickhouse_connect; ..."