Project Name: metricsinfra
Type of quota increase requested: 8 CPU / 16G ram available
Reason: The single Prometheus instance is using all of the projects RAM quota and half of the CPUs. This quota increase would be used to experiment with and build out related services required for more wide-spread use:
- Local puppet master to manage secrets and patches still in development
- Separate alertmanager instance for ease of scaling and HA
- Experiment with Thanos and/or Cortex for scaling up Prometheus to support all Cloud VPS instances
- Build tools to allow projects to self-manage their data sources and alerting rules
If possible I'd like to try out the Bullseye pre-releases too - these instances likely won't be permanent and being already able to use newer software versions in a fast-moving platform is useful.