Benchmarks of Web Servers and SQL Drivers

Web Servers

• go net/http
• fastapi
• express.js
• next.js
• spring boot + webflux

SQL Drivers

• github.com/lib/pq
• github.com/jackc/pgx
• asyncpg
• psycopg
• node-postgres
• porsager/postgres
• postgresql r2dbc

Details

• hardware: Apple Silicon M4 (10 cores) 32GB RAM
• runtime: Docker Desktop 4.53 (8 cores, 24GB RAM, 4GB swap)
• database engine: Postgres 17.5 (Docker)
• database pooler: pgcat 0.2.5
• load tester: grafana/k6
• load test duration: 3m sustained
• load success threshold: error rate < 0.01

Top Results

Rank	Web Server + SQL Driver	VUs	RPS[K]	avg[ms]	p90[ms]
1	go net/http + jackc/pgx	130	9.1	4	8
2	spring boot + webflux + postgresql r2dbc	100	7.7	3	5
3	express.js + node-postgres	100	5.8	7	9
4	express.js + porsager/postgres	100	5.2	9	13
5	fastapi + asyncpg	130	3.7	24	31
6	fastapi + psycopg	100	3.0	22	36
7	fastapi + psycopg + pgcat	100	2.8	35	51
8	next.js + porsager/postgres	100	2.6	37	43
9	go net/http + lib/pq w/o p. stmt	100	1.8	43	87

Results (Full)

Web Server	SQL Driver	VUs	RPS[K]	min[ms]	max[ms]	avg[ms]	p90[ms]	p95[ms]
go net/http	lib/pq w/o p. stmt	100	1.8	1.2	510	43	87	115
	jackc/pgx w/o p. stmt	100	7.1	.79	169	3	6	9
	jackc/pgx	100	8.6	.21	128	1	2	13
		130	9.1	.30	114	4	8	11
		500	6.3	.67	2.4s	68	166	252
		1K	5.7	57	3.7s	176	273	362
	jackc/pgx + pgcat	100	7.9	10	243	12	14	15
		130	8.4	.69	77	5	10	15
		250	7.0	10	.9s	35	72	99
		500	6.6	14	1s	64	105	135
		2K	6.5	233	1.3s	305	357	388
		4K	6.5	61	2.1s	.6s	.7s	.7s
		6K	6.9	151	2.3s	.8s	1s	1s
		10K	5.7	451	3.7s	1.7s	2.1s	2.4s
fastapi	asyncpg	100	3.7	11	156	16	20	23
		130	3.7	2	356	24	31	35
		200	3.2	16	.8s	51	66	75
	psycopg	100	3.0	2.3	272	22	36	45
		200	2.7	14	1.6s	62	141	191
	psycopg + pgcat	100	2.8	14	328	35	51	62
		200	2.4	15	1.5s	69	157	203
express.js	node-postgres	100	5.8	.83	291	7	9	10
		250	4.8	4.0	1.2s	51	61	66
	porsager/postgres	100	5.2	.43	329	9	13	15
		250	4.9	4.3	1.1s	41	48	51
next.js	porsager/postgres	100	2.6	4.1	543	37	43	47
		250	2.4	8.5	3.3s	94	112	119
spring boot + we bflux	postgresql r2dbc	100	7.7	.29	46	3	5	7
		130	7.5	1.9	81	7	10	13
		250	7.7	3.2	83	33	26	27
		500	7.0	27	192	60	70	77
		1K	6.8	31	250	136	153	160
	postgresql r2dbc + pgcat	10K	4.4	59	3.6s	2.2s	2.7s	2.8s

Setup

How To Run Benchmarks

1. run docker compose build to build all services (you may need to start by configuring Docker resources)
2. run docker compose up -d postgres to start database
3. run docker compose up service_name (check compose.yml for list of services)
4. configure benchmark.js options (vux, duration, or stages)
5. run k6 run benchmark.js --env URL=http://localhost:8080 (for next.js, use http://localhost:8080/api/tx)

Workload

The workload consists in performing the following SQL transactions. We quickly conclude the execution of prepared statement by database engine is most efficient. Performance difference is shown in results for Go net/http + jackc/pgx.

SQL

BEGIN;
  UPDATE pgbench_accounts 
    SET abalance = abalance + :delta 
    WHERE aid = :aid; 
  SELECT abalance 
    FROM pgbench_accounts 
    WHERE aid = :aid;
  UPDATE pgbench_tellers 
    SET tbalance = tbalance + :delta 
    WHERE tid = :tid; 
  UPDATE pgbench_branches 
    SET bbalance = bbalance + :delta 
    WHERE bid = :bid;
  INSERT INTO pgbench_history (tid, bid, aid, delta, mtime, filler) 
    VALUES (:tid, :bid, :aid, :delta, now(), repeat('x',22));
END;

Schema

pgbench_accounts

aid	integer
bid	integer
abalance	integer
filler	character(84)

pgbench_branches

bid	integer
bbalance	integer
filler	character(88)

pgbench_history

tid	integer
bid	integer
aid	integer
delta	integer
mtime	timestamp without time zone
filler	character(22)

pgbench_tellers

tid	integer
bid	integer
tbalance	integer
filler	character(84)

Notes

• pgcat does not support asyncpg
• pgcat does not support node-postgres
• pgcat does not support porsager/postgres