With respect to: https://github.com/SigNoz/signoz-otel-collector/tree/main/exporter/clickhousetracesexporter/migrations Given that there can be multiple otel-collector processes running on a cluster, what prevents the race condition of several of those processes trying to perform the migration schema changes when an upgrade to signoz is rolled out? Sometimes I find that the otel-collector processes enter crash-loops because some thing the schema is on V3 and some think it's on V10 and it seems risky that mutliple processes would try to rollout the schema change