-
-
Notifications
You must be signed in to change notification settings - Fork 503
Infrastructure & Ops
Meg Gutshall edited this page Dec 28, 2020
·
9 revisions
- Single server hosted on Azure sponsored hosting with a staging server on Linode.
- Rails
- PostgresDB
- Puma
- Instance is used for all Diaper Banks (hosted service, not federated or self-hosted)
- We're in production out of beta!
- These diaper banks are trying to use the system with real-world data
- That means the system is IN PRODUCTION and is the source-of-truth for some users
- There are no particular times of usage, though likely US-daylight
- Q: How do we communicate with users, like if we needed downtime?
- Testing users are encouraged to submit GitHub issues (all of them? trained?)
- TravisCI deploys master
- Triggered by merge into master -- now that we're in production, auto deploys go to our staging server (diaperbase.org.) Deploying to production requires tagging a release and then manually deploying it with Capistrano.
- Runs Capistrano deploy if the rest of the tests pass
- See .travis.yml
- Capistrano is used to perform the deploy
- See config/deploy.rb for general config
- See config/deploy/production.rb for production config
- Single server for DB and web server (Puma)
- Uses hard-coded ip address
- We have the ability to deploy from a branch
cap production deploy BRANCH=branch_to_deploy
- Bugsnag (we have an OSS account, contact an admin for access to it)
- Skylight - Monitoring service
- A few admins (a-a-ron, sean, jcavena) have production shell access
- We have the
capistrano-rails-console
gem installed. This allows us to runcap production rails:console
which opens a production rails console from your local command line.
- Issue #405 will add DB backups
- VM Snapshot (?)