tose down / restart

Stop deployments or restart pods.

tose down

Stop the current deployment.

Synopsis

bash

tose down [project] [-y]

Description

Stops the project. If the project has a running deployment, tose down stops that one, even when a newer build is in progress. This scales the app to zero and cancels the project's builds that have not reached deploy (pending, cloning, building, pushing). With nothing running, it only cancels the latest build.

After the stop, the CLI re-reads the deployment and the project's deployments and reports the final status. It never prints stopped for a stop that did not take effect. See What Happens When You Stop and Exit codes.

Arguments

Argument	Description
`[project]`	Optional. Project slug or `workspace/project` format

Options

Option	Alias	Description
`--yes`	`-y`	Skip confirmation prompt. Required without a terminal or with `--json`

Usage

Interactive stop with confirmation:

bash

$ tose down

Stop deployment for my-workspace/my-app? No
? Confirm: Yes

✓ Project stopped
  Deployment dep_abc123 is now stopped.
  Builds of this project that had not reached deploy were cancelled too.

Stop specific project:

bash

$ tose down my-workspace/my-app

Stop deployment for my-workspace/my-app? Yes
✓ Project stopped
  Deployment dep_abc123 is now stopped.
  Builds of this project that had not reached deploy were cancelled too.

Skip confirmation:

bash

$ tose down -y

✓ Project stopped
  Deployment dep_abc123 is now stopped.
  Builds of this project that had not reached deploy were cancelled too.

Project is busy (the CLI retries):

bash

$ tose down -y

⠹ Project is busy; retrying the stop in 12s (attempt 1/4)...
✓ Project stopped

JSON Output

bash

$ tose down -y --json

json

{
  "success": true,
  "data": {
    "id": "dep_abc123",
    "status": "stopped"
  }
}

What Happens When You Stop

Application receives a graceful shutdown signal
Active connections are drained
Containers are terminated
Deployment status changes to stopped

Stopping a running deployment stops the whole project. In one locked write, every deployment of the project that is running or deploying, and every one still before deploy (pending, cloning, building, pushing), is recorded as stopped. The project stays at zero replicas until you deploy again.

Only deployments created before the stop began are affected. A deploy requested after the stop builds and ships normally.
A stopped build gets no deploy, no email and no success or failure event.
Stopping a deploying deployment finishes asynchronously. The builder puts back the version that served last, if one exists, otherwise it scales the workload to 0. The deployment can end running again, with the reason in error_message, if its own image is still the one serving. Run tose status after a stop to see the final state.
If the project's workload lock cannot be taken within 60 seconds, the API answers 503 Service Unavailable with a Retry-After header and nothing is stopped. See Retries and timeouts.
A later redeploy never picks the image of a cancelled rollout.

See Stop Deployment in the API reference.

Retries and timeouts

503 with Retry-After: nothing was stopped. The CLI waits for the Retry-After time (at most 60 seconds, or 5 seconds when the header is absent) and sends the Stop request again, up to 4 attempts in total.
Still busy after 4 attempts: the command fails with OPERATION_FAILED (exit code 8, retryable: true). The message says the stop did not happen; run tose down again in a minute.
Timeout: the CLI waits up to about 100 seconds for the Stop request. If that runs out, it fails with TIMEOUT (exit code 7, retryable: true) and the outcome is marked unknown (details.stop_outcome: "unknown"). The server may still have finished the stop, so run tose status before retrying.

Final status

After the Stop request is accepted, the CLI re-reads the deployment (waiting up to about 30 seconds while it is still deploying) and then the project's deployments:

The deployment is stopped and no deployment of the project is running: success, exit code 0.
The deployment, or another deployment of the project, ended running again, for example because a rollout's own image was still serving: the command fails with OPERATION_FAILED (exit code 8) and shows the deployment's error_message. It does not claim stopped.
The deployment is still deploying after the wait: the command fails with TIMEOUT (exit code 7). Run tose status to see the final state.

Exit codes

Code	Name	When
`0`	`SUCCESS`	The project is stopped, or the only build was cancelled
`4`	`CONFIRMATION_REQUIRED`	No terminal, or `--json`, without `--yes`. Nothing is sent
`5`	`CANCELLED`	You declined the prompt
`6`	`NOT_FOUND`	The project has no deployments
`7`	`TIMEOUT`	The Stop request timed out (outcome unknown), or the deployment is still `deploying` after the stop
`8`	`OPERATION_FAILED`	The project stayed busy after 4 attempts, or the stop ended with a deployment `running` again

The latest deployment already being stopped, failed or otherwise finished is not an error: the CLI warns and exits 0 without sending a request. See Exit codes for the full list.

tose deployment cancel <id> uses the same retry and final-status check. See tose deployment cancel.

tose restart

Restart pods without redeploying.

Synopsis

bash

tose restart [project]

Description

Restarts all running pods for the project. This does not trigger a new deployment or rebuild - it simply restarts the existing containers. Useful for recovering from transient issues or applying config changes that only require a restart.

Arguments

Argument	Description
`[project]`	Optional. Project slug or `workspace/project` format

Usage

bash

$ tose restart

✓ Restarting my-workspace/my-app...
✓ Restart initiated
  5 pod(s) restarting

Restart specific project:

bash

$ tose restart my-workspace/my-app

✓ Restarting my-workspace/my-app...
✓ Restart initiated
  3 pod(s) restarting

JSON Output

bash

$ tose restart --json

json

{
  "success": true,
  "data": {
    "restarted_pods": 5
  }
}

down vs restart vs redeploy

Action	What it does	Use when
`tose down`	Stops deployment completely	Maintenance, freeing resources
`tose restart`	Restarts pods (same image)	Transient errors, memory leaks
`tose redeploy`	Redeploys (same image, new pods)	Config changes
`tose deploy`	Full build + deploy	Code changes

Recovery After Stop

To bring a stopped deployment back up:

bash

# Redeploy (no rebuild)
tose redeploy

# Or full deploy (rebuild from latest commit)
tose deploy

tose status - Check if deployment is stopped
tose deployment cancel - Cancel one build by ID
tose deploy - Deploy after stopping
tose redeploy - Redeploy without rebuilding
tose logs - View logs before stopping

tose down / restart ​

tose down ​

Synopsis ​

Description ​

Arguments ​

Options ​

Usage ​

JSON Output ​

What Happens When You Stop ​

Retries and timeouts ​

Final status ​

Exit codes ​

tose restart ​

Synopsis ​

Description ​

Arguments ​

Usage ​

JSON Output ​

down vs restart vs redeploy ​

Recovery After Stop ​

Related Commands ​

tose down / restart

tose down

Synopsis

Description

Arguments

Options

Usage

JSON Output

What Happens When You Stop

Retries and timeouts

Final status

Exit codes

tose restart

Synopsis

Description

Arguments

Usage

JSON Output

down vs restart vs redeploy

Recovery After Stop

Related Commands