- Write splitting of the mongodb data vol: I did not know this but mongo production notes include a section about multiple volumes - 1 for data + indexes, another for the journal and a final one for logs to maximize iops potential of a single vm. Also discussing are piops which can be helpful too. (At work we got a bit more performance out of our data volume by provisioning more iops last week. Note: This only works up to the point of your next bottleneck which for us was hit fairly quickly afterwards. eg 1k iops more made room for more work, a further 1k iops didn’t help beyond the first.)
- Caddy, nginx performance eval: Nice comparison of caddy vs nginx for static file serving and proxy type workloads. Describes methodology and results for the benchmark.
- Why Open Source Firmware is Important: (Video) I did not realize there was so much going on beneath the kernel. I mean yeah there’s the h/w boot process which is software but the current state of the art there seems to be bloated and unsafe. Open source firmware sounds like a good idea. How do you make it happen industry wide. Sounds like a hard problem.
