I’m trying out the regional deployment with a Rails app hitting Postgres. Starting with a single region near me (ams) with both database and web, all was working fine. Then I added a second region (lax). Naturally, the requests hitting LAX responded much slower as they had to send database traffic back to AMS. So a few questions here.
After a second deployment of the web app, it got deployed in SJC and LAX, SJC showing up as a backup of LAX. Shouldn’t the second deployment be in AMS or a backup of it?
Instances ID VERSION REGION DESIRED STATUS HEALTH CHECKS RESTARTS CREATED be1f688e 17 lax run running 1 total, 1 passing 0 19m34s ago 6059bef1 17 sjc(B) run running 1 total, 1 passing 0 20m13s ago Lebrija:ensayo joshua$ fly regions list Region Pool: ams lax Backup Region: fra lhr sea sjc
Also, what would be the winning strategy for the LAX instance accessing postgres? If I setup an LAX replica, will the default behavior be to read from the replica and write to the master in AMS? Forgive my laziness looking up how stolon operates.