Builds consistently terminated with exit code 137 after ~25 minutes

I'm trying to build the official local Bot API server as Cloud Function using the Node.js 26 runtime.

My Function build spec is set to 4 CPU / 4 GB RAM.

The project successfully configures and compiles, but the build keeps getting terminated with:

Build failed with exit code 137.

I've tested different build parallelism:

-j4 → ~72%
-j2 → ~81%
-j1 → ~55%

The failures occur at roughly 25 minutes into the build.

I then added continuous resource monitoring. The container reports:

Mem: 15.6G total

but the cgroup reports:

memory.max = 4294967296

which is a 4 GiB cgroup limit.

More importantly, immediately before one of the failures, the container still had:

memory.current ≈ 573 MB
oom = 0
oom_kill = 0

and approximately:

212 GB free disk

The build was still actively compiling when it was terminated. There are no compiler errors; it simply ends with status code 137.

Build failed with exit code 137.

My main questions are:

1. Is there any ~25-minute build/process limit on Appwrite Cloud that could terminate the build with SIGKILL/137?
2. Can Appwrite's build infrastructure return 137 for reasons other than an OOM kill?
3. Is there any way to see the actual reason/process that caused the SIGKILL?

My understanding is that Pro/Scale builds have a longer documented timeout(of about 45 minutes), so I'm trying to understand why this consistently dies at around 25 minutes.

I can provide the full verbose build log if needed.

full build log

anyone? I tried so many times