Add promotion-time settings flags to model deployment promote by mingminglang-baseten · Pull Request #176 · basetenlabs/baseten-cli

mingminglang-baseten · 2026-10-21T02:05:28Z

🚀 What

baseten model deployment promote can change the target environment's settings as part of the promotion (region, instance type, request backpressure, autoscaling, and autoscaling schedules as raw API JSON), and override promotion settings for that promotion only (--override-*). Autoscaling and override flags get their own help sections, and the confirmation prompt names the settings that will change. A promote without the new flags sends the same body as before.

💻 How

Fills environment_settings_update and promotion_settings_override, sending each group only when one of its flags is passed.
The autoscaling flags are the existing shared ones, now in their own help section, so update-autoscaling shows that heading too.

🔬 Testing

New promote tests for each settings group and for rejected input: --instance-type with --preserve-env-instance-type, and empty or trailing schedule JSON (the server rejects unknown keys itself). go mod tidy, go fmt, go vet and the model tests pass.

🤖 Generated with Claude Code

…listing The usage command listed your teams and your Code settings in each with GET /v1/routes/settings/users/me/teams, which is being removed. Use GET /v1/teams for your teams and GET /v1/routes/settings/teams/{team}/users/me for your settings in each, which say the same things. Org admins are members of every team, so they still see every team. Claude-Session: https://claude.ai/code/session_01UkWMh7GwDgbmR88spBWMEm

Pins baseten-go to the client regenerated for the renamed Grid API, so every call site moves to the GetGrid*/PostGrid*/PatchGrid*/DeleteGrid* methods. grid usage now reads your settings from GET /v1/grid/team_settings/users/me and passes --team straight through as team_id: the server uses your only team, or asks for --team when you belong to several. The client-side team listing and name matching are gone, along with the "No usage this month" branch they fed. Claude-Session: https://claude.ai/code/session_01UkWMh7GwDgbmR88spBWMEm

Preserve the source PR's all-team usage and team name/ID filtering while using the current Grid API and Grid-named files. Resolve the usage implementation and test conflicts without restoring legacy command files.

* adding functionality to cli * Update go.sum * adding functionality to cli

`grid usage` listed your teams and fetched your settings in each to draw your spend against your limit, the one place the CLI still stitched teams together client-side. Per the review on this PR, usage is one command: `grid route usage [--team] [--user-id]`, whose arguments behave as the API does. Without --team, usage spans the teams you can see; without --user-id, members see their own usage and admins see everyone's. Your spend against your limit in a team is `grid user-settings describe`. Claude-Session: https://claude.ai/code/session_01UkWMh7GwDgbmR88spBWMEm

The list spans every team you can see unless --team narrows it, so two rows can share a name. A TEAM column, as the connection list already has, tells them apart. Claude-Session: https://claude.ai/code/session_01UkWMh7GwDgbmR88spBWMEm

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

…nges in the prompt Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

Copilot

🟡 Changes recommended

Schedule validation can bypass explicitly empty input and nested misspellings, permitting promotion without intended settings changes.

2 open issues

What changed in this PR

Extends deployment promotion to update environment settings and apply one-time promotion overrides, reusing existing settings mappings.

Changes:

Adds grouped flags, help examples, and settings-aware confirmation.
Builds optional settings payloads and adds validation tests.
Updates the Go SDK and adapts promotion settings types.

File	Description
internal/cmd/command.model_environment.go	Adapts SDK types and shared mappings.
internal/cmd/command.model_deployment.go	Builds promotion payloads and validates schedule JSON.
internal/cmd/command.model_deployment_test.go	Tests settings payloads and rejected inputs.
go.sum	Updates SDK checksums.
go.mod	Updates the SDK dependency.
cmd/command.model.go	Defines reusable settings flag groups.
cmd/command.model_environment.go	Embeds shared promotion flags.
cmd/command.model_deployment.go	Adds promotion flags, help, and examples.

_{Review effort: Balanced}

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

talyaacovi

lgtm!

cretz

Looks great, only minor things (most non-blocking)

cretz · 2026-10-21T16:55:59Z

 	PreserveEnvInstanceType bool   `flag:"preserve-env-instance-type" desc:"Keep the target environment's current instance type instead of applying the promoted deployment's."`

+	Region                    OptionalFlag[string] `flag:"region" desc:"Region slug to set on the environment. Pass an empty string for Global. Run 'baseten org regions' to see the slugs available."`
+	InstanceTypeID            OptionalFlag[string] `flag:"instance-type-id" desc:"Instance type to set on the environment. Cannot be combined with --preserve-env-instance-type."`


This is rough on users, we don't expose instance type IDs and we don't expect users to have to provide IDs. When you create a model for example, you provide the instance type name (I know we had discussed this internally during API review, though I forget where we landed). I wonder if the API can't change, if we should accept --instance-type here and go look up the ID (unsure if that replaces or is in addition to this).

This seems to be more of a naming issue I think

When you create a model for example, you provide the instance type name

What user provides in truss is what the REST API calls id (e.g. "A10Gx4x16" in truss https://docs.baseten.co/reference/truss-configuration#resources). The name in REST looks like "L4:4x16 - 1 L4 GPU, 24 GiB video memory, 4 vCPU, 16 GiB" (I don't think users would want to pass this). In addition, instance_type in REST is the full struct including id, name and some other fields.

I can update this to --instance-type to align with truss convention. For reference, is it ok to use docs since we don't have a baseten model instance-types command yet?

CPU: https://docs.baseten.co/deployment/resources#cpu-only-instances
GPU: https://docs.baseten.co/deployment/resources#gpu-instances

Oh interesting, yes, I think we should update the flag name to --instance-type. Yes, we can just point to docs (though would totally support baseten model instance-types in a followup).

cretz · 2026-10-21T17:47:15Z

+// --autoscaling-* prefix, for commands that set more than autoscaling. Its
+// fields mirror [AutoscalingSettingsFlags] so one converts to the other.
+type PrefixedAutoscalingSettingsFlags struct {
+	MinReplica                  OptionalFlag[int] `flag:"autoscaling-min-replica" desc:"Minimum number of replicas." group:"autoscaling" group-pri:"200"`


Not a requirement, but IMO the field name matching the flag name is ideal (even if it is an annoying prefix everywhere, but out users suffer that too)

Yeah I'll keep the field name and flag name aligned. On the naming, I'm thinking maybe we should remove the prefix as we discussed in the other thread on push-time settings.

Arguments for prefixing:

Precedence of prefixing (admittedly this is an example prefixing makes lots of sense): https://github.com/basetenlabs/baseten-cli/blob/main/cmd/command.volume_sync.go#L152-L158

We don't need to worry about colliding field names under different parent fields

Structurally analogous to how we handle request backpressure policy in the existing environment update command, as --autoscaling-min-replica is similar to --request-backpressure-policy; --min-replica is similar to --policy (the latter sounds a bit too general?)

baseten model environment update-autoscaling ... --min-replica 2 baseten model environment update-request-backpressure ... --policy reject-on-full

(That said, if the same logic is applied to promotion setting, --promotion-rolling-deploy or --override-promotion-rolling-deploy seems verbose, so maybe we can treat --request-backpressure-policy as a special case.)

Arguments for not prefixing:

Simplicity

Users might be used to --min-replica etc flag that already exists for update

I concur with removing the prefix, and we won't have it on model push either looking at #180

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

cretz

LGTM, I do concur with making the flag name --instance-type and removing the --autoscaling- prefix on promote flags

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

youngzheng-oss and others added 23 commits October 15, 2026 09:42

grid rename

11378b5

Update command.harness_usage_test.go

2f35d1b

Merge branch 'main' into shivain/grid-rename

12e56e3

refactor(grid): align command files and symbols with Grid hierarchy

c518c93

Merge PR #165 usage team discovery into Grid commands

7397675

Preserve the source PR's all-team usage and team name/ID filtering while using the current Grid API and Grid-named files. Resolve the usage implementation and test conflicts without restoring legacy command files.

Update go.sum

0ee24ff

feat(grid): Add settings, connections, and classifier commands (#173)

83efd56

* adding functionality to cli * Update go.sum * adding functionality to cli

responding to comments (1)

9f45d0d

responding to comments (2)

208984d

renaming

a52e787

Update command.grid.go

d3f8bf5

Removing to match spec

1caded8

updating with new sdk

a827c61

grid settings flag rename to route id

19b14d8

updated create/update outputs to be concise confirmations

0125e94

Update command.grid_test.go

f26ba07

test renaming

bf1a1f2

matching create (use hasGridRouteClassifier)

a4802ba

Show the team in grid route list

0de0e7f

The list spans every team you can see unless --team narrows it, so two rows can share a name. A TEAM column, as the connection list already has, tells them apart. Claude-Session: https://claude.ai/code/session_01UkWMh7GwDgbmR88spBWMEm

Add promotion-time settings flags to model deployment promote

ed5413a

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

mingminglang-baseten marked this pull request as ready for review October 21, 2026 02:09

Reject conflicting promote flags, tighten schedule JSON, and name cha…

81572b8

…nges in the prompt Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

mingminglang-baseten requested a balanced review from Copilot October 21, 2026 02:30

Copilot started reviewing on behalf of mingminglang-baseten October 21, 2026 02:30

View session

Copilot AI reviewed Oct 21, 2026

View reviewed changes

Comment thread internal/cmd/command.model_deployment.go

Comment thread internal/cmd/command.model_deployment.go

Reject an empty schedule update and misspelled keys inside schedules

76757d4

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

talyaacovi approved these changes Oct 21, 2026

View reviewed changes

talyaacovi mentioned this pull request Oct 21, 2026

Add hidden traffic-rule commands for environment shadow rules #179

Open

cretz reviewed Oct 21, 2026

View reviewed changes

Base automatically changed from shivain/grid-rename to main October 21, 2026 19:58

mingminglang-baseten and others added 2 commits October 21, 2026 21:04

Merge origin/main into mingming/promotion-time-settings

be28ce8

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

Inline promote flags and leave schedule validation to the server

8727802

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

mingminglang-baseten requested a review from cretz October 22, 2026 02:31

mingminglang-baseten mentioned this pull request Oct 22, 2026

Add deployment and promotion settings flags to model push #180

Open

cretz reviewed Oct 22, 2026

View reviewed changes

Rename promote flags to --instance-type and unprefixed autoscaling flags

d6d5138

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

Conversation

mingminglang-baseten commented Oct 21, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🚀 What

💻 How

🔬 Testing

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

🟡 Changes recommended

Uh oh!

Uh oh!

Uh oh!

talyaacovi left a comment

Choose a reason for hiding this comment

Uh oh!

cretz left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

cretz Oct 21, 2026

Choose a reason for hiding this comment

Uh oh!

mingminglang-baseten Oct 22, 2026

Choose a reason for hiding this comment

Uh oh!

cretz Oct 22, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

cretz Oct 21, 2026

Choose a reason for hiding this comment

Uh oh!

mingminglang-baseten Oct 22, 2026

Choose a reason for hiding this comment

Uh oh!

cretz Oct 22, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

cretz left a comment

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

mingminglang-baseten commented Oct 21, 2026 •

edited

Loading

cretz Oct 22, 2026 •

edited

Loading