CLI reference
The @putnami/cloud extension adds 20 commands under putnami cloud.
putnami cloud --help lists them, together with the machine commands and the
older names described at the end of this page. putnami cloud <command> help
lists the verbs of one command. The Key flags column is selective: run
putnami cloud <command> --help for every flag of the installed command.
Commands that return records support --output=jsonl for scripts and agents.
| Command | What it covers |
|---|---|
login, logout, whoami |
Your session on this machine |
setup |
Link this repository to a workspace |
status |
One line per entry below that has a status: ok, degraded, failing or unknown |
token |
A bearer for a script, and workspace machine tokens |
registries |
The npm, Go, OCI and put registries on this machine |
packages |
Namespaces, read grants, public mirrors, image moves, archive publication |
channels |
Move a channel, read it, or follow one |
ci |
Hosted CI runs, recovery, and the putnami.ci.json contract |
cache |
The remote build cache |
source |
The GitHub repository linked to the workspace |
config, secrets |
Project configuration and secrets |
db |
Databases: access grants, a local bridge, migrations |
env |
Environments: what runs where, prerequisites, convergence |
deploy |
Show one release, or submit an operator release request |
logs, traces, metrics |
Query a project's telemetry |
Status
putnami cloud status prints one line per entry that has a status: its state,
what it found, and the next command to run. The next command is the entry's
fix when the entry is not ok and names one, else the command that shows the
entry in full. For example, 7 of the 12 lines:
STATE ENTRY DETAIL NEXT
ok whoami you@example.com, workspace ws-acme putnami cloud whoami
ok registries npm, go, oci on demand; put key active putnami cloud registries status
degraded packages npm, go not needed; oci not bound; put bound putnami cloud packages status
ok ci 9 of 9 checks healthy; 439 runs, ~€8.79 in 7 days putnami cloud ci status
ok config 12 projects, every required key set, 40 optional missing putnami cloud config status
degraded db 4 databases, 18 bindings ready, 1 not bound by a selected workload putnami cloud db status
failing env prod: 11 of 12 workloads ready, canary gen 315 putnami cloud deploy status cm_721cd93aThe 12 entries are whoami, token, registries, packages, channels,
ci, cache, source, config, secrets, db and env. Each line is the
header of that entry's own status command, described in
Status of each entry.
The command reads the entries in parallel and covers the whole workspace,
wherever you run it. An entry whose read fails, or that does not answer within
15 seconds, reads unknown with the reason, and every other line still prints.
Until you are signed in and the repository is linked, every line except
whoami and registries reads unknown. The env line covers every
environment putnami.ci.json declares; --env <env> limits it to one.
The command exits 1 when an entry is failing, and 0 otherwise. With
--strict, it also exits 1 when an entry is degraded or unknown: use that
form to gate CI. With --output=json, data is {state, entries}: state is
the worst entry, and entries holds the full node of each entry, metrics and
checks included.
The deployment table that putnami cloud status printed before is
putnami cloud env status --health or putnami cloud env status --provenance.
Status of each entry
Every entry that has a state has its own status command:
putnami cloud <entry> status [<name>] [--strict] [--output=json]<name>, where the table below shows it, narrows the status to one item, such
as one project for config status or one database for db status.
putnami cloud whoami takes
no status verb: it is its own status. putnami cloud deploy status <release-id> and putnami cloud env doctor [<env>] print the same shape.
The output starts with a header line: the title, the state, and what the entry
found. A table of metrics follows, then a table of checks. A check that is not
ok prints its fix under it. When the entry is not ok and no check names a
fix, a last line next: <command> names the entry's own fix:
packages degraded npm, go not needed; oci not bound; put bound
METRIC VALUE WINDOW
bindings active 1 now
STATE CHECK DETAIL
ok npm not needed
ok go not needed
degraded oci not bound; putnami.ci.json declares environments
fix: putnami cloud packages namespaces activate --protocol oci --namespace acme --idempotency-key acme-oci
ok put namespace acmeEvery status uses the same 4 states:
| State | Means |
|---|---|
ok |
What is declared exists and works. |
degraded |
The entry works, with a gap that needs attention. |
failing |
The entry does not work. |
unknown |
The entry did not answer, or a check could not run. |
An entry takes the state of its worst check. From best to worst, the order is
ok, degraded, unknown, failing.
Every status command exits by the same rule:
- It exits 1 when the state is
failing, and 0 otherwise. - With
--strict, it also exits 1 when the state isdegradedorunknown. - A
failingstatus also prints<title> is <state>: <detail>on standard error. - A command that cannot start still exits with its own code: 2 for a usage error, 3 for a session the service refuses.
With --output=json, the command prints one result envelope, and its data
is the node, on success and on failure alike. A node has these fields; an
empty field is left out:
| Field | Holds |
|---|---|
id |
A stable id for scripts, such as ci or ci.runner |
title |
The name a person reads |
state |
ok, degraded, failing or unknown |
detail |
What the check found |
fix |
The command that clears a state other than ok, when one exists |
metrics |
The measured values, described below |
facts |
The identifiers a script reads and detail only quotes, such as an image digest or a run id |
children |
The checks the state comes from, each a node of the same shape |
A metric has an id, a title, a value and a unit: count, bytes,
percent, seconds or eur. limit is the total or the quota the value
reads against, when one applies. window is the period a counted value
covers, such as 7 days or last run; without it, the value is measured now.
estimated is true for a value computed from rates instead of measured.
kind is usage or health:
- A
usagemetric is a quantity the workspace consumes, the kind a bill or a quota reads: runs, spend, keys set. - A
healthmetric measures how well the entry works: queued runs, ready workloads, cache reuse.
The data of putnami cloud registries status --output=json, trimmed:
{
"id": "registries",
"title": "registries",
"state": "ok",
"detail": "npm, go, oci on demand; put key active",
"metrics": [
{"id": "registries_configured", "title": "registries configured", "value": 4, "unit": "count", "limit": 4, "kind": "health"}
],
"children": [
{"id": "registries.npm", "title": "npm", "state": "ok", "detail": "on demand, npm.putnami.dev"}
]
}Read one value with jq:
putnami cloud ci status --output=json | jq '.data.metrics[] | select(.id=="queued_runs") | .value'| Command | Checks | Usage metrics | Health metrics | Fails when |
|---|---|---|---|---|
whoami |
The session on this machine, and the workspace this repository is linked to | session expires in | You are not signed in, or the repository is not linked | |
token status |
Each workspace machine token that is not revoked | active machine tokens | machine tokens expiring within 7 days | Never. A token past its expiry, or expiring within 7 days, is degraded |
registries status [<registry>] |
Each registry endpoint on this machine, and the state of its stored key | registries configured | A stored key is revoked or expired | |
packages status [<protocol>] |
The npm, go, oci and put namespace bindings this checkout needs. Without an oci binding, oci is ok when another workspace granted yours at least one active OCI publisher package share |
bindings active | Never. A needed binding that is missing is degraded |
|
channels status [<channel>] |
Each channel putnami.ci.json declares, and the generation each registry applied |
channels declared | registries behind | A registry is still behind at the --wait deadline |
ci status |
Each CI readiness check, including the open pull request heads that have no run, then the dispatch, the live run and the last run as facts | runs, spend (7 days) | queued runs, stuck runs, oldest queued, undelivered checks, reused tasks (last run) | A run is stuck, the runner is stale, or check results wait to be published. A pull request head that has no run, other than a draft's, is degraded |
cache status |
.putnami/cache.json, and the identity its token source signs in as |
reused tasks, served from cache, executed (last run) | Never | |
source status |
The connected GitHub repository, its GitHub App installation, and the App's permissions | open pull requests | stale pull requests, with no update in 14 days | The installation is suspended or removed, or the repository left it |
config status [<project>] |
Each project's declared config keys against what each environment resolves | projects, keys declared, keys set | required missing, optional missing | A required key has no value and no default |
secrets status [<project>] |
Each project's declared secrets, set or not. It never reads a value | secrets set | required missing, optional missing | A required secret is not set |
db status [<db>] |
Each database, the workloads putnami.ci.json selects that bind it, and each binding that is not ready |
databases, schemas; with <db> and an active grant, size and tables |
bindings ready, databases not bound; with <db> and an active grant, connections |
Never. A binding that is not ready, or a database no selected workload binds, is degraded |
env status [<env>] |
Each environment putnami.ci.json declares: the newest move of its channel, and each workload declared or deployed |
workloads declared | workloads ready, workloads not deployed, workloads not declared, CD generation | A workload failed, or ended partial or skipped |
config status and secrets status without a project stop after 240 service
calls, so that putnami cloud status answers in time. The projects past that
budget form one unknown check named not read; run the status of each one by
name. The config apps whose name no project of the checkout has, such as the
old name of a renamed project, form one degraded check named no project.
Their keys do not count. An app that a service account reads through an active
secret grant, such as a shared GitHub App key, is not in that check: the fact
shared_apps names it. The status reads the grants only when an app has no
project. Listing them needs platform.workspace.manage; when the list is
refused, the no project check is unknown, because a shared app may be in
it.
putnami cloud config status shop/workloads/api lists, under each
environment, each key that has no value and no default:
config shop/workloads/api failing 1 required key missing
METRIC VALUE WINDOW
keys declared 12 now
keys set 11 of 12 (92%) now
required missing 1 now
optional missing 0 now
STATE CHECK DETAIL
failing prod 1 required key missing: stripe.endpoint
fix: putnami cloud config publish shop/workloads/api --env prod
failing stripe.endpoint required, no value and no defaultA key that resolves to its schema default counts as set. The fix is
config publish, which writes the values your committed conf/env*.yaml
files hold.
putnami cloud secrets status shop/workloads/api lists each secret that is
not set, with the command that sets it from standard input:
secrets shop/workloads/api failing 1 required secret missing in prod
METRIC VALUE WINDOW
secrets set 2 of 3 (67%) now
required missing 1 now
optional missing 0 now
STATE CHECK DETAIL
failing prod 1 required secret missing: stripe.api_key
failing stripe.api_key required, not set
fix: putnami cloud secrets set shop/workloads/api stripe.api_key --env prod --from-stdinputnami cloud db status <db> shows one database. It never opens a grant: the
size, tables and connections show only while you hold an active grant on the
database. Without one, a metrics check names the db grant request command.
db shop_orders ok postgres, 2 schemas, bound by 2 workloads, your read grant active
METRIC VALUE WINDOW
size 212 MiB now
tables 38 now
connections 4 now
STATE CHECK DETAIL
ok schemas 2 schemas: migration, orders
ok shop/workloads/api binds it as orders in prod
ok shop/workloads/worker binds it as orders in prod
ok grant your read grant is active until 2026-10-04T12:00:00ZDeploy and release
A merge to main publishes a release set and moves a channel. The platform
then deploys every environment that follows that channel. See
Environments follow channels.
| Command | Does | Key flags |
|---|---|---|
putnami publish |
Build and publish selected project artifacts, config, migrations, archives, deployment declarations, and site-content members | --env, --archives, --namespace, --dry-run, --skip-publish-config, --skip-publish-migration, --skip-publish-deployment |
putnami cloud deploy status <release-id> |
Show one release as a status: one check per workload. Exits 1 when the release failed or ended partial or skipped; a release still converging reads degraded |
--strict, --output=json |
putnami cloud deploy publish-v2 --request-file <path> |
Needs the workspace deploy permission. Submit an exact release set to one environment for the projects the request names. To roll one project back, name an older release set | --request-file, --wait |
putnami cloud env status [<env>] |
Show each environment putnami.ci.json declares, or the one you name, against what it runs: its channel and each workload. --health and --provenance print the deployment table instead |
--env, --workspace, --strict, --provenance, --health, --output=json |
putnami cloud env doctor [<env>] |
Check every prerequisite continuous delivery (CD) needs for one environment and print the fix for each one that is not met. Read-only; exits 1 when a prerequisite is missing, which reads failing. A check that could not run reads unknown and fails the command only with --strict |
--env, --strict, --workspace, --output=json |
putnami cloud env enable [<env>] |
Apply every workspace-owned prerequisite of one environment (namespace bindings, environment definition, Runtime grants), report the operator-only rows with their fix, then print the doctor table. A second run on the same commit writes nothing; exits 1 when a step failed or a row is still missing |
--env, --source-revision, --workspace, --output=json |
putnami cloud env status covers every environment putnami.ci.json
declares, then every other environment where the workspace runs a workload.
Under each environment, it checks the channel the environment follows, each
workload putnami.ci.json declares, and each workload that runs there without
a declaration. This sample leaves out the 11 workloads that read ok:
env failing prod: 11 of 12 workloads ready, canary gen 315
METRIC VALUE WINDOW
workloads declared 12 now
workloads ready 11 of 12 (92%) now
workloads not deployed 0 now
workloads not declared 0 now
prod CD generation 315 now
STATE CHECK DETAIL
failing prod 11 of 12 workloads ready, canary gen 315
ok channel canary gen 315 opened 2026-10-04T08:18Z (rs_8fa9d34e…), release cm_721cd93a
failing shop/workloads/web failed, revision web-00041: image pull refused
fix: putnami cloud deploy status cm_721cd93aA workload that is ready reads ok. These read degraded:
- A workload still deploying.
- A workload that serves, but whose last run failed.
- A workload declared and not deployed.
- A workload deployed and not declared.
A workload that failed, or whose release ended partial or skipped, reads
failing, and the command exits 1.
With --output=json, data is the status node. --health and --provenance
keep the deployment JSON: workspace_id, deployments, channel_follow, and
with --health, health. A script that reads deployments passes one of them.
putnami cloud env status --health and --provenance read the environment you
name, else the environment .putnami/cloud-link.json records, else prod.
--health prints the deployment table: one row per project and environment,
with its serving revision, its health, and its last run. --provenance prints
the provenance table instead. Both start with one CD line for the
environment they read. The line names every channel the environment follows and
what the newest move of that channel did:
CD prod ← canary: gen 42 converged 2026-09-18T10:02Z (rs_ab12cd34…)A move that deployed nothing or only part of the environment also shows its
reason, and a move Control is still delivering reads in flight. When the header
cannot be read, the line reads CD prod: unavailable (<reason>) and the table
still prints. Structured output carries the same data under channel_follow.
The LAST RUN, TRIGGER, and REASON columns describe each project's most
recent run in its environment, failed runs included, while HEALTH keeps
describing the last successful release. They are read from the 100 most recent
runs of each environment; a project whose last run is older shows -.
TRIGGER reads cli for a run no channel move opened, and unknown when
Control could not read its channel-move records.
The release and commit of a row describe the last deploy that made the project
serve, whether a channel move or an operator request opened it. The
revision and deploy time describe that deploy too, unless a config write rolled
a newer revision since: the row then shows that revision and when it rolled. A
deploy that failed after its revision took the traffic counts as the last
deploy, and STATE shows the failure.
A deploy that a channel move opened is read from the same 100 most recent
runs: a project whose last such deploy is older shows its last deploy that
no channel move opened instead. CONFIG VERSION reads - for a deploy that a
channel move opened, because that deploy pins no config version.
putnami cloud env status --provenance prints one row per project:
PROJECT, COMMIT, TREE, RELEASE, REVISION, and CONFIG VERSION.
COMMIT and TREE show the first 7 characters. TREE is the git tree of the
commit's content, recorded by the channel move that opened the project's last
run. It shows only when that run deployed the commit in COMMIT, so the two
always describe the same source. It reads - when the run was on another
commit, when no channel move opened the run, or when the move carried no tree.
Structured output carries the full tree as trigger.source_tree.
putnami cloud env status prod --health answers "is prod on main and
healthy". It adds three columns to the rows of the --env environment. Rows of
other environments show - in them.
| Column | Shows |
|---|---|
ERRORS (10m) |
The number of log entries at severity ERROR or above in the last 10 minutes |
TOP ERROR |
The most frequent of those messages: the message or msg field of a JSON entry, else the text, cut to its first line |
ON MAIN |
yes, no, or unknown: whether the served commit is on main in your local checkout |
The errors come from one logs read for the whole environment, not one read per
service. The read stops after three pages of up to 1,000 entries each. When
more entries remain, every count is a lower bound: the cell reads N+ (for
example 12+), a note line above the table says so, and structured output sets
health.logs.capped. Entries with no service label, or
with a label that matches no row, are counted in a note line and never added to
a row. When the logs cannot be read (a machine token, which cannot read logs,
an older control plane, or a backend error), the note reads
Errors prod: unavailable (<reason>), the error cells show -, and the command
still succeeds.
ON MAIN compares the served commit with origin/main, or main when
origin/main does not exist, in your local checkout. It never fetches, so run
git fetch first for a current answer. It reads yes when the commit is on
main and no when your checkout has the commit and it is not on main. A pull
request head that was squash-merged reads no: that exact commit is not on
main. It reads unknown when no commit is recorded, the directory is not a git
checkout, there is no main ref, or your checkout does not have the commit.
Structured output carries every field, including the reason for each
unknown, under the health key; without --health that key is absent.
putnami cloud env doctor prints the status of one environment's
prerequisites, with one check per prerequisite. A prerequisite that is met
reads ok. One that is missing reads failing, with its fix. One that could
not be checked reads unknown, as every server check does when the control
plane does not answer. The command exits 1 when a prerequisite is missing;
with --strict it also exits 1 on an unknown check, which is the form to use
as a CI gate. Before the first deploy
lists the checks and the ones that can read unknown.
putnami cloud env enable applies every fix the workspace owns among those
prerequisites and reports the rest, then prints the prerequisites as a table: ok,
missing or unknown on each row. Each write reads the current
revision first and sends it as the expected revision, so a second run writes
nothing. It needs platform.workspace.manage; a session without it is refused
before any write. Enable the environment in one command
lists what each row gets.
Identity and workspace
| Command | Does | Key flags |
|---|---|---|
putnami cloud login |
Sign in with OAuth device authorization | --no-open, --client-id, --scope |
putnami cloud logout |
Clear local Cloud credentials and revoke reachable registry tokens | |
putnami cloud whoami |
Show the account signed in on this machine and the workspace the repository is linked to. Exits 1 when either is missing | --strict, --output=json |
putnami cloud setup |
Link the repository to a workspace. With no linked workspace, create one; --slug chooses its slug |
--workspace, --workspace-name, --slug, --environment, --no-cache, --auto |
putnami intelligence setup |
Enable Intelligence, install its agent wiring, and report discovered hosted Intelligence availability | --disable, --graph |
putnami cloud token [<target>] |
Mint a workspace bearer for a script, or a bearer for one target | cache, npm, go, oci or put (or --for <target>), --materialize |
putnami cloud token create|list|revoke |
Create, list, or revoke workspace-owned machine tokens | --name, --scopes |
putnami cloud token status |
Check each machine token of the workspace that is not revoked. One past its expiry, or expiring within 7 days, reads degraded |
--strict, --output=json |
putnami cloud source |
Connect, inspect, or disconnect the workspace GitHub repository. status checks the connection, the GitHub App installation, and its permissions |
connect, status, disconnect |
Hosted CI and channels
| Command | Does | Key flags |
|---|---|---|
putnami cloud ci |
Enable the subscription; inspect, start, cancel, retry, and diagnose runs; operate recovery modes | enable|disable|subscription, list|view|logs, start|cancel|retry, status|usage|insights|timings, repair|pause|drain|resume |
putnami cloud ci report |
Submit a delivery-run conclusion for the current pushed commit | --conclusion, --exit-code, or -- <cmd…>; optional --status |
putnami cloud ci init|validate|fmt|explain |
Write, check, format, or explain putnami.ci.json |
init --force, fmt --check, explain --event push|tag|pull_request --branch <b>|--git-tag <t> [--pr <n>] |
putnami cloud channels set <channel> |
Point a channel at a release set that already exists: promotion and rollback | --from <channel|rs_id>, --expected <rs_id> |
putnami cloud channels status [<channel>] |
Show the head of a channel and the generation each registry applied; without a name, every channel putnami.ci.json declares |
--wait <duration>, --strict, --output=json |
putnami cloud channels follow <namespace> <channel|rs_id> |
Make the workspace's CI follow a producer channel, or pin one release set, optionally changing the runner channel at the same time | --runner-channel |
Run IDs may be shortened to any unambiguous prefix. CI mutations require
--reason; pass --yes as well in a non-interactive environment. See
Hosted CI for the setup and recovery path.
putnami cloud ci init writes a putnami.ci.json that runs the lint,
test, build and validate jobs your extensions declare, and nothing else:
no publication and no environment. validate checks the file against its
schema and against the jobs your extensions declare. fmt rewrites it in
canonical form, and fmt --check only reports. explain shows which rule an
event matches and what the run would publish, without calling the provider.
putnami cloud channels set moves a channel only if its head is still the one
the command read, so a concurrent move makes it fail instead of being lost.
putnami cloud channels status reads degraded while a registry is behind, and
exits 0: registries apply a move a few seconds after it. --wait 5m polls
until every registry serves the head. A registry still behind when the time
runs out reads failing, and the command exits 1.
Configuration and data access
| Command | Does | Key flags |
|---|---|---|
putnami cloud config status [<project>] |
Check each project's declared config keys against what each environment resolves. Exits 1 when a required key has no value and no default | --strict, --output=json |
putnami cloud config show [<project>] |
Show the resolved config, or one view of it | --declared, --schema, --key, --secret-keys, --reveal-secrets, --keys, --metadata, --env, --with-secrets |
putnami cloud config validate [<project>] |
Validate local authored Config inputs without packaging or network access | --env, --schema-from, --namespace |
putnami cloud config drift <project> |
Compare the committed conf/env*.yaml with the published config; exits non-zero on drift |
|
putnami cloud config put <project> |
Force-write non-secret values from a YAML file | --config-from |
putnami cloud secrets status [<project>] |
Check each project's declared secrets in each environment. Never reads a value. Exits 1 when a required secret is not set | --strict, --output=json |
putnami cloud secrets |
Manage workspace and project secrets | --block, --field, --from-stdin, --from-file, --reveal, --yes |
putnami cloud db status [<db>] |
Check each database against the workloads putnami.ci.json selects. Never opens a grant |
--strict, --output=json |
putnami cloud db |
Inspect databases, manage time-boxed access, or open a local bridge | list, info <db>, grant request|list|revoke, connect, --database, --level, --reason, --ttl |
Observability and delivery evidence
| Command | Does | Key flags |
|---|---|---|
putnami cloud logs |
Query or live-tail application logs | --app, --since, --level, --follow, --cursor |
putnami cloud metrics |
Query application metric series | --app, --window, --from, --to, --cursor |
putnami cloud traces |
Query traces and filter slow requests | --app, --slow, --from, --to, --cursor |
putnami intelligence audit-attest |
Get, set, or batch-prefetch immutable Intelligence audit attestations | get|set|prefetch, --repo, --project, --group, --input-hash, --rubric-hash |
putnami intelligence provenance |
Record or list signed provenance receipts: which agent run took part in a commit | record|show, --commit, --repository, --family, --model, --run, --run-kind |
putnami intelligence review |
Manage workspace subscription reviews, owner credentials, and workers | doctor|status|launch|pause|resume|reconcile|cancel, auth, worker |
putnami intelligence provenance record records one receipt for one commit, HEAD
by default. A review labels its independence: how sure the service is that the
reviewer model differs from the author models. The label is full only when
every commit of the reviewed branch holds a verified receipt. One commit
without a receipt lowers it to partial. Record a receipt for every commit of
the branch:
for c in $(git rev-list <base>..HEAD); do putnami intelligence provenance record --family <family> --model <model-id> --commit "$c"; doneThe complete intelligence review command families are:
| Family | Verbs |
|---|---|
| Review control | doctor, status, launch, pause, resume, reconcile, cancel |
| Owner credentials | auth login, auth replace, auth status, auth revoke for codex or claude-code |
| Review workers | worker enroll, worker list, worker revoke, worker run |
Registries, packages and cache
| Command | Does | Key flags |
|---|---|---|
putnami cloud registries |
Configure registry endpoints, check them and their stored keys, or revoke legacy credentials | setup, status [<registry>] (also the default), logout, optional --registry |
putnami cloud packages status [<protocol>] |
Check that the linked workspace binds each npm, go, oci and put namespace this checkout needs, or publishes its OCI images through package shares another workspace granted it | --strict, --output=json |
putnami cloud packages namespaces|grants|mirrors |
Manage namespace bindings, read grants, and public mirror targets | namespaces activate|list, grants create|list|revoke, mirrors add|list|remove|rotate-credential; mirrors add takes --ecosystem npm|oci|archive, --id, --destination, optional --alias, and --package for archive |
putnami cloud packages copy|retag|revert |
Copy, retag, or revert OCI images server-side | copy <src> <dst>, retag <ref> <tag…>, revert <repo>:<channel> |
putnami cloud cache |
Check or disable the remote build cache. status also shows how many tasks the last CI run took from the cache |
status (also the default), disable, optional --remove |
Direct publishers
putnami publish runs these for you. Run one directly to publish a single
input again.
| Command | Does | Key flags |
|---|---|---|
putnami cloud config publish |
Register an app schema and publish conf/env.<env>.yaml |
--app, --env, --dry-run |
putnami cloud db migrations publish |
Publish a workload migration bundle | --app, --bundle-from, --namespace, --bundle-version, --dry-run |
putnami cloud packages publish |
Publish pre-built extension archives as one immutable Put version | --app, --archives-from, --binary-name, --dry-run |
Machine commands
These commands serve the engine, publication and image packaging. Prefer
putnami publish and putnami package for the normal project workflow.
| Command | Does | Key flags |
|---|---|---|
putnami cloud release-set |
Serve exact immutable release-set provider operations from a private request document | resolve|release|channel-set|channel-status, --request-file <absolute-mode-0600-file> |
putnami cloud image-layers |
Produce or verify the deterministic tarballs declared by an image project's image-layers.json |
--project, --check |
putnami cloud registry-token |
Mint your registry bearer for one known host | --host, --materialize |
Older command names
The names below still run. Use the new form in scripts and docs.
| Older form | Use instead |
|---|---|
putnami cloud status for the deployment table |
putnami cloud env status --health or putnami cloud env status --provenance |
putnami cloud tokens create|list|revoke |
putnami cloud token create|list|revoke |
putnami cloud distribution bindings … |
putnami cloud packages namespaces … |
putnami cloud distribution grants|mirrors … |
putnami cloud packages grants|mirrors … |
putnami cloud oci copy|retag|revert |
putnami cloud packages copy|retag|revert |
putnami cloud publish-archives |
putnami cloud packages publish |
putnami cloud track |
putnami cloud channels follow |
putnami cloud report |
putnami cloud ci report |
putnami cloud validate-config |
putnami cloud config validate |
putnami cloud publish-config |
putnami cloud config publish |
putnami cloud publish-migration |
putnami cloud db migrations publish |
putnami cloud config <project>, config list, config resolve |
putnami cloud config show <project>, show --keys, show --metadata |
putnami channel set|status |
putnami cloud channels set|status |
putnami ci init|validate|fmt|explain |
putnami cloud ci init|validate|fmt|explain |
putnami cloud status now prints the one-line summary, and refuses --health
and --provenance with exit 2. A script that parsed its deployment table must
call putnami cloud env status --health or putnami cloud env status --provenance. Plain putnami cloud env status prints the status node.