pier_

Pause and resume

Stop a resource, an environment, a project or a workspace and keep everything — and what happens when the platform pauses you

Pausing stops the running parts of something and keeps its data, configuration, hostnames and history. Resuming brings it back as it was. The verbs are the same at every scope.

pier pause api
output
OK  Paused app "api"
→  [22:36:07] active
pier ls apps
output
                                                                                       
 NAME STATUS REPLICAS AUTOSCALE IMAGE                                    PORT UPDATED  
                                                                                       
 api  paused 1/1      -         zot.pier-registry.svc.cluster.local:5... 8080 just now 
                                                                                       
pier resume api
output
OK  Resumed app "api"
→  [22:36:07] resumed

Scopes

CommandPauses
pier pause api, pier pause api/canary, pier pause main, pier pause wwwone resource or one target
pier environments pause stagingevery resource in the environment
pier projects pause acmeevery environment of the project
pier workspaces pauseeverything in the workspace

A paused app keeps its hostname, which answers with an error page; a paused Postgres keeps its data and stops accepting connections; a paused site is not served. Buckets and volumes pause with their environment. A deploy into a paused target offers to resume it first (--no-resume fails instead, for CI).

A pause requested with pier pause does not expire: it is recorded with the reason user_paused, and only pier resume clears it. A paused app is not billed for compute; its storage is billed.

When the platform pauses you

The platform pauses a whole workspace for three reasons. The reason is shown in the console, in pier billing overview, and in an email:

ReasonTriggerRecovery windowResumes when
cap_hitoverage spend reached the workspace capnever purged; held storage is billedthe cap is raised or the period rolls over
payment_failedan invoice is seven days past due30 days, then purgedthe invoice is paid
trial_expiredthe trial's grant or its 14 days ran out7 days, then purgeda plan is chosen

A system pause cannot be lifted with pier resume; it is resolved through billing. pier events --all-projects --severity warn shows the enforcement events.

Removed resources

pier rm is a pause with a 30-day clock: the resource leaves the environment, its data is kept, pier trash lists it, pier trash restore <type> <name> brings it back (config-external until re-declared in pier.yaml), pier trash purge ends the window. Concepts: trash.

See also

pier pause · pier resume · pier trash · Workspaces and billing

On this page