My Always Free Oracle instance got idle-reclaimed, and I discovered my quota silently dropped from 4 OCPU/24GB to 2 OCPU/12GB
I've been running my app's backend (Node.js + MongoDB) on an Oracle Always Free Ampere A1 instance (4 OCPU/24GB) for months. It got automatically terminated by Oracle's idle-reclamation policy after a quiet week of low traffic.
No big deal, I thought — I still had my boot volume (47GB, fully intact), so I just needed to spin up a new instance from it.
Except when I tried, I hit LimitExceeded errors. Turns out my tenancy's free-tier quota had been silently cut to 2 OCPU/12GB — half of what I had before — following Oracle's June 2026 policy change to the Always Free tier.
Then even at the reduced 2/12 spec, I hit "Out of host capacity" — Mumbai region simply has no free ARM hosts available right now.
So in one evening: my working production server was gone, my quota was cut in half without warning, and even the reduced allocation couldn't be provisioned due to capacity. All while trying to launch a feature I'd been building for weeks.
Data's safe (boot volume survived), but the whole experience was rough — especially not knowing this quota change had even happened until I hit the wall trying to recover.
Anyone else run into this? Considering PAYG or just moving to a paid VPS at this point.