Efficient use of skunk sessions

[wvandermerwe]

Currently the session pool is passed to the methods implementing the query/command logic. After doing some reading in the skunk Discord channel I found a few discussions mentioning that this would be more efficient:

object Main extends ResourceApp.Forever {

override def run(args: List[String]): Resource[IO, Unit] = {
for {
sessionPool <- Session.pooled[IO](
host = host,
port = port,
user = user,
database = db,
password = password,
max = max
)
databaseAccessLayer = sessionPool.map(new DBAccessLayerIO)
service = new MyServiceIO
server <- BlazeServerBuilder[IO]
.withHttpApp(service.orNotFound)
.bindHttp(host = "0.0.0.0", port = 8080)
.resource
.void
} yield server

From what I understand this would mean "using" the dbAccessLayer on each request inside the route, so I comes down to what is more costly, inefficient session reuse or repeatedly allocating new objects? I'm not sure which one would be the better tradeoff to make in a "typical" web backend.

https://discord.com/channels/632277896739946517/839844929877704785/955838009872384020