Downtime costs roughly $10,000 perminute in a hyperscale and is categorised as the highest risk.
This is pretty wild quote. I wonder what the source is?
Downtime costs roughly $10,000 perminute in a hyperscale and is categorised as the highest risk.
This is pretty wild quote. I wonder what the source is?
If you need to ensure migrations run in a certain order with regular db:migrate, set up Outrigger.ordered. It can be a hash or a proc that takes a tag; either way it needs to return a sortable value: Outrigger.ordered = { predeploy: -1, postdeploy: 1 } This will run predeploys, untagged migrations (implicitly 0), and then postdeploy migrations.
class PreDeployMigration < ActiveRecord::Migration tag :predeploy end
This is especially useful for zero downtime deploys to Production environments.
A transition period rather than Stop-The-World migration; we want to merge in a few repositories per day, with minimal disruption to work-flow.