Item: PostgreSQL
Rating: 10
Author: Paul De Audney

Use Cases and Deployment Scope

PostgreSQL is used across a wide number of systems. Ranging from customer-facing primary data storage of traditional relational data to using it more like a NoSQL data store with the JSON & JSONB data types. Analytical workloads in some parts of the business are serviced by PostgreSQL as well.

Pros and Cons

Permissive licensing
JSONB data types allow for migration from NoSQL data stores that haven't scaled as well as would have like when providing consistency guarantees.
Various index types to support full text search.
Extensions & customization of the database.
Geo-spatial support with routing support.

Default tuning isn't optimized for modern hardware
No native support for multi-master setups
Scaling out with partitions across multiple servers can create transaction issues in some scenarios

Most Important Features

GIN & GIST indexes for text search.
pg_repack extension to support online table bloat clean up.
Upcoming btree deduplication.
Query analysis using auto_explain for slow queries.

Return on Investment

Performance over NoSQL solutions when you introduce data consistency guarantees
Licensing costs don't exist

Alternatives Considered

MongoDB, MySQL and MariaDB Platform

We selected PostgreSQL due to the number of employees who have used it in the past.
The data consistency guarantees.
The multiple transaction isolation levels support.
The performance tuning options available.
The query performance analysis capabilities.
The long history of continual improvements made to the database.
The very strong and supportive opensource community around PostgreSQL.

Key Insights

Do you think PostgreSQL delivers good value for the price?

Yes

Are you happy with PostgreSQL's feature set?

Yes

Did PostgreSQL live up to sales and marketing promises?

Yes

Did implementation of PostgreSQL go as expected?

Yes

Would you buy PostgreSQL again?

Yes

Other Software Used

Amazon DynamoDB, MySQL

Likelihood to Recommend

PostgreSQL is generally well suited to basically any database workload one can think of.

JSONB data types are great for dealing with various use cases that come up to avoid an EAV pattern.
Custom data types can be supported.
Various extensions can really add a lot of excellent features.
Logical replication in later versions supports per table replication.

NoSQL workloads shine on PostgreSQL

Overall Satisfaction with PostgreSQL