I got called to come to my son's school to help change his insulin pod. I got there to discover that the app on his phone had thrown a memory corruption error and wiped out all of his insulin dosage settings
Fortunately, I have been keeping a copy of those settings in a self-hosted git repository, mostly so that I could look through the git history and see how the numbers change over time. This helps me understand the impacts that some of these changes are having
So there I am in his school's office on my phone reading a CSV off of my @forgejo instance and plugging all of his dosage settings back into the app on his phone
We use these tools to store our source code. But here I am using it for the code that keeps my son alive.
#t1d
I've got plans to further improve the JVM's behaviour in container environments, hopefully I'll carve out time to work on them in the near future. The problem is partially accounting for other processes (e.g. readiness probes, liveness probes) that might be occasionally launched in the same cgroup, so we can't eat 100% of the memory limit. Gauging what % is tricky