Anonymizing Data in Postgres Without Losing Your Mind (or Your
summary
Shows concrete options and tradeoffs: an in-DB extension that does dynamic masking and anonymized dumps but no replication, a CLI dumper that applies YAML rules for batched exports, and an open‑source pipeline called pgstream that performs fast snapshots, supports logical replication and DDL tracking, streams transformed rows to Postgres/Elasticsearch/Kafka, integrates security‑label rules, handles JSONB with templating, and runs in production for very large DBs. Benchmarks claim better throughput than pg_dump/restore; limits include extra infra, no XML support, and no automatic sensitive‑data heuristics.
AI-written summary · about
chapters (8)
- 0:00Introduction by hosts
- 0:21Defining the data problem
- 3:18Challenges of anonymization
- 5:13Plain Postgres anonymization methods
- 8:59External anonymization tools overview
- 13:16Deep dive into PG stream
- 22:50Key takeaways and recommendations
- 24:17Q&A session
related talks
32mAnonymization and Dynamic Masking with PostgreSQL
51mWhen Postgres Can't, You Can!
28mOpen the gate a little: strategies to protect and share data
34mDatabase Anonymization with PostgreSQL
04mData Masking in PG
29mFinally Respecting GDPR Article 25 with PostgreSQL Anonymizer
25mBig Bad World of Postgres Dev Environments
37mPostgres at Any Scale