There are a number of "roles" in the Socorro infra. The resource profiles for these roles are not identical, ergo there will be different AWS instance types for different roles.
The roles are:
- Admin
- Collectors
- Crash-Analysis
- Elasticsearch
- Middleware
- Postgres
- Processors
- RabbitMQ
- SymbolAPI
- Webheads
Admin nodes are responsible for orchestrating deployments and running Crontabber. For historical reasons beyond the scope of this document, there are three admin nodes in PHX1; however, only two (one each for Stage and Prod) will be required in AWS.
The orchestration of deployments is relatively lightweight, requiring only enough resources to pull some repositories, unpack some tarballs, that sort of thing. The various Crontabber jobs are also fairly lightweight (from the perspective of the Admin node, at least), since they're mostly just triggering actions on other nodes (largely DB-related).
Note that current admin node sp-admin01 (a Xeon server with two processors and 8 GB of RAM) is pinned constantly. This is only because we're running a bunch of migration-related tasks on it that won't be a factor once we're in AWS for realsies.