pg_probackup vs Quemsi

pg_probackup is a free utility for backing up and recovering PostgreSQL database clusters. It supports full backups and page-level incremental DELTA, PAGE, and PTRACK modes, along with incremental and partial restores. Point-in-time recovery uses WAL archives. The utility can check data consistency automatically, validate backups on demand, and verify PostgreSQL instances with checkdb. Retention policies can use recovery time, backup count, or a time-to-live value; expired backups can be merged or deleted. Backup, restore, merge, delete, validation, and checkdb tasks can run on multiple threads. It supports remote PostgreSQL instances and backups from standby servers. Postgres Pro Enterprise edition supports S3 storage through MinIO, Amazon S3, and VK Cloud. The repository lists compatibility with PostgreSQL 13 through 18. Linux packages and Windows installers are available. Remote backup over SSH is not supported on Windows, and SSH operations require matching local and remote pg_probackup versions. The license permits use, copying, modification, and distribution without a fee, but disclaims maintenance, support, updates, and warranties.

Quemsi is a hybrid tool for database backup and restore, with a web UI for tagged snapshots, timeline history, and one-click restoration. It supports MySQL, PostgreSQL, SQL Server, and MongoDB. Backup and restore flows can include custom steps such as stopping services or running scripts, and backups can go to AWS S3, Azure Blob, or local drives under the user's control. Its agent runs operations inside the user's infrastructure, while the cloud service handles metadata rather than database content. The agent is open source and can be audited, modified, or built from source. Agent-to-cloud communication uses TLS 1.3, agents use certificate-based mutual TLS, and SHA-256 checksums verify backup integrity. Agents are available for Windows, Linux, and macOS, and can also run as a Java JAR or Docker container. The free Solo plan costs nothing and includes unlimited databases, environments, and flows for one user, with community support.