Performance is an important goal for pgagroal and effort have been made to make pgagroal scale and use a limited number of resources.
This report describe pgagroal in relationship to 3 other PostgreSQL
connection pool implementations, which we will call a, b and c.
The pgbench program was used in the runs. All pool configurations were made with performance in mind.
All diagrams are using the same identifier for the connection pool in question, so a is a in all
diagrams and so on.
The runs were performed on RHEL 7.7 /
EPEL / DevTools 8
based machines on 10G network. All connection pools were the latest versions as of January 14, 2020. pgagroal was
using the epoll mode of libev.
This run uses
pgbench -M simple
This run uses
pgbench -M extended
This run uses
pgbench -M prepared
This run uses
pgbench -S -M prepared
Please, run your own benchmarks to see how pgagroal compare to your existing connection pool deployment.