AWS provisioning

A 1 GB EC2 instance (t3.micro, t4g.micro) with no swap runs Synthigy, but the engine is its biggest process, about 500 MB. When anything else on the box asks for memory that is not there — apt's daily package refresh, a busy proxy — the kernel's OOM killer kills the engine. Swap, one less memory spike and a service that relaunches the engine keep the box up.

Step 3 needs CLI v0.1.11 or newer (synthigy version; synthigy upgrade).

# 0. check: swap should show 0
free -h

# 1. 2 GB swap
sudo fallocate -l 2G /swapfile && sudo chmod 600 /swapfile
sudo mkswap /swapfile && sudo swapon /swapfile
echo '/swapfile none swap sw 0 0' | sudo tee -a /etc/fstab
echo 'vm.swappiness=10' | sudo tee /etc/sysctl.d/99-swap.conf && sudo sysctl --system

# 2. remove the apt hook that ran appstreamcli (not needed on a server)
apt-get -s purge appstream        # dry run: should remove only appstream
sudo apt-get purge -y appstream

# 3. run Synthigy as a service so it starts at boot and outlives logout
cd ~/myapp                        # the folder that holds .synthigy/
synthigy down
synthigy service install
sudo loginctl enable-linger $USER
synthigy service status

Checking it

• free -h shows 2.0Gi of swap.
• synthigy service status shows active (running).
• When the engine is killed, the portal relaunches it and says so in the journal:

journalctl -k | grep -i oom            # what the kernel killed, and when
journalctl --user -u synthigy -f       # Engine died: killed by SIGKILL, ... — relaunching.

An engine that was up is relaunched each time it dies. One that dies while starting is not: the portal keeps running and shows why (synthigy logs).

See Running Synthigy for services in general.