The `ps:scale` command now accepts a `--format` flag that defaults to `stdout` and can be set to `json` to emit the current formation as a JSON array of process type and quantity objects, matching the JSON output the `:report` subcommands already provide. The flag only applies when displaying the current formation; it is ignored when process types are supplied for scaling. When no scale has been set for the app, the JSON output is an empty array.
The `apps:list` command now accepts a `--format` flag that defaults to `stdout` and can be set to `json` to emit the app names as a JSON array, matching the JSON output the `:report` subcommands already provide. When no apps exist, the JSON output is an empty array.
For an undeployed app, the additive diff covers every chart resource, so
the literal 'has been added' header appears once per resource - six times
in the python test app's chart. `assert_output_contains` defaults to
exact-count 1, which fails. Pass -1 (at least 1) to match the actual
semantics of the test.
Adds a `scheduler-k3s:preview <app>` command that renders a unified diff
between the manifests currently stored in the live Helm release for an app
and the manifests that the next deploy would roll out. The diff is produced
entirely via the Helm Go SDK that scheduler-k3s already uses, with no
external `helm` binary or `helm-diff` plugin install required. The
underlying diff and parse logic is vendored from `databus23/helm-diff`
(Apache 2.0) into a new internal package, trimmed of three-way-merge,
release-ownership tracking, and the non-default output formatters.
The chart-construction half of `TriggerSchedulerDeploy` has been extracted
into a reusable `BuildAppChart` helper so deploy and preview share the
same chart materialization without duplicating template logic. Secret
values are redacted by default in preview output; `--show-secrets` and
`--show-secrets-decoded` flags match the upstream `helm diff` UX. Only the
main app helm release is compared in this iteration; auxiliary releases
for config vars, image-pull secrets, and TLS certificates are out of scope
and documented as such.
Annotation and label values containing a newline previously round-tripped
as two adjacent `key: value` lines, and the second line then failed the
`SplitN(line, ": ", 2)` parse on read. Move the per-`(processType,
resourceType)` files to `PropertyMap*` storage so `\n` in values and `/`
in keys are preserved verbatim. An idempotent `TriggerInstall` migration
walks `--global` and every app, converting any legacy line-formatted file
in place by probing `PropertyMapGet` first and rewriting via
`PropertyListGet` only when the probe fails. Property names are
unchanged, so the annotations/labels report scanner and the
`reservedAnnotationPrefixes` filter keep working without modification.
Closes#8719.
Chart property names containing `/` failed because per-key flat-file storage interpreted the slash as a filesystem path separator, and any move to line-based storage would silently truncate multi-line values. A new `PropertyMap*` helper family in `common` persists each map as a single JSON file so both `/` in keys and `\n` in values round-trip losslessly. Chart overrides move to `chart-overrides.<chart>` and an idempotent `TriggerInstall` migration rewrites any legacy `chart.<chart>.<key>` files into the new layout. The deprecated `scheduler-k3s:set --global chart.*` form is rerouted through the same map storage so it does not silently orphan writes. Closes#8717.
The scheduler-k3s plugin exposes `annotations:set` and `labels:set` but no matching report subcommands, so scripts could not inspect the configured state for idempotent management without reading the property files directly. The new `scheduler-k3s:annotations:report` and `scheduler-k3s:labels:report` surface the configured entries with stdout, JSON, single-flag query, and `--process-type` / `--resource-type` filtering, mirroring the recently added `scheduler-k3s:charts:report`. The literal `GlobalProcessType` value `--global` is rendered as `global` in report keys to avoid leading dashes. The `scheduler-k3s:autoscaling-auth:report` command is updated alongside to loop over every app when no app and no `--global` flag is provided, matching the convention used by `scheduler-k3s:report`.
Helm chart overrides on the scheduler-k3s plugin were previously set
through the generic `scheduler-k3s:set --global chart.<chart>.<property>`
interface, which mixed chart-level configuration into the same command
that manages scheduler properties and offered no focused way to inspect
which overrides were configured. A dedicated `scheduler-k3s:charts:set`
sets and clears chart-specific helm values, and a complementary
`scheduler-k3s:charts:report` surfaces configured overrides per chart
with optional JSON output and single-field flag queries. The legacy form
on `scheduler-k3s:set` continues to work but now emits a deprecation
warning pointing users at the new subcommand.
storage:report gains a flat dotted key per attachment field (entry-name, host-path, container-path, phases, process-type, subpath, readonly, volume-options, volume-chown) under the --storage-attachment.<index>.<field> shape. Both stdout and JSON pick the keys up through the existing common.ReportSingleApp pipeline. Attachments that reference a missing storage entry emit a warning and are skipped, so the rest of the report still renders. Closes#8710.
Re-running `storage:mount <app> <entry> --container-dir <path>` against an existing `(entry, container_dir, process_type)` tuple now updates the attachment's mount-time fields in place instead of erroring with `already mounted`. Declarative tooling that wants to change `--volume-options`, `--volume-chown`, `--phase`, `--volume-subpath`, or `--volume-readonly` on an existing mount no longer has to unmount-then-remount, which briefly dropped the volume from `storage:report` and raced against any deploy that fired in the window. The legacy `host:container[:opts]` form keeps its strict `Mount path already exists.` failure.
The `storage:list <app> --format json` payload conflated the
attachment's `Readonly` flag and its `VolumeOptions` field into a
single derived `volume_options` string that rendered as `ro`,
`<options>`, or `ro,<options>` depending on which fields were set on
the underlying attachment. That shape is fine for the legacy
`host:container[:options]` text view but it leaves drift-detection
tooling unable to tell whether a `ro` token came from
`Attachment.Readonly == true` or from the operator setting
`Attachment.VolumeOptions = "ro"` directly. The JSON now exposes
`readonly` (boolean) and `volume_options` (string) as separate
omitempty keys populated straight from the attachment, and
`formatStorageListEntry` combines them at format time for the colon-
form text view. `ParseMountPath` was extended in lockstep so callers
of the legacy `host:container:opts` form no longer have to special-
case the `ro` token themselves.
The named-entry form of `storage:mount` had no way to set the attachment's `volume_options`, even though the legacy `host:container:opts` colon form parses options into the same field and every downstream consumer (the docker-args trigger and the storage:list/storage:report display path) already renders them. Tooling that declaratively re-applies attachments silently dropped options on every re-apply.
The apps plugin accepted `apps:set --global deploy-source` and
`apps:set --global deploy-source-metadata` even though nothing reads
those properties globally, accepted `apps:set <app> disable-autocreation`
even though only the global form is consulted by `maybeCreateApp`, and
never emitted `disable-autocreation` in `apps:report` for either scope.
Drop the global `deploy-source*` writes from `GlobalProperties`, reject
per-app `disable-autocreation` writes at the plugin level, and surface
`--app-global-disable-autocreation` in both the per-app and `--global`
reports so the property the runtime actually reads is round-trippable.
storage:create --chown invoked chown-storage-dir with the full host path, but the helper validates a basename and prepends the storage root itself, so every call failed with `Directory can only contain the following set of characters`. Pass the entry name and reject the combination of --chown with a non-default host path, since the helper only manages the default storage location.
`storage:destroy` now prompts for confirmation before removing a named storage entry, matching the behavior of other destructive commands such as `apps:destroy` and `network:destroy`. The prompt can be skipped with the `--force` flag or the global `dokku --force` flag for non-interactive callers.
Emitting the restart policy as a `--restart` argument from a `docker-args-process-deploy` trigger exposed a latent gluing bug: proxy and builder triggers concatenated `$STDIN$output` with no separator, and the scheduler and builders appended `docker-args-process-*` output directly onto the older `docker-args-*` output, so adjacent arguments could merge into values like `--restart=on-failure:10--label`. A space is now inserted at both the trigger echo and the concatenation so arguments always stay separated.
`dokku ps:set <app> restart-policy` with no value erroneously returned `Invalid restart-policy specified` instead of unsetting the property like every other ps property. The restart policy is now managed as a normal app and global property surfaced through the `--ps-restart-policy`, `--ps-global-restart-policy`, and `--ps-computed-restart-policy` report flags, with the effective value applied at deploy time and existing values migrated on install. Because it is no longer stored as a Docker option it no longer appears in `docker-options:report`, and `--ps-restart-policy` now reports the raw value with the `on-failure:10` default available via `--ps-computed-restart-policy`.
Exposes raw, global, and computed report flags for every openresty per-app property and allows those properties to be set with `--global`, so external tooling can distinguish a property that was never set from one left at its built-in default. The computed value resolves the per-app value first, then the global value, then the default. Also corrects `client-header-timeout` which read the `client-body-timeout` key and sets the computed `client-max-body-size` default to `1m`.
Every Go-implemented plugin's `:report --format json` now emits keys without the redundant `<plugin>-` head segment, matching the shape bash plugins have always emitted. The CLI flag names and `:set` semantics are unchanged. For backwards compatibility during the 0.38.x patch series, the legacy `<plugin>-<property>` keys are emitted side-by-side with the new keys, and a future major release will drop the legacy keys. `common.ReportSingleApp` is refactored to accept a `ReportSingleAppInput` struct with a `Validate()` method so the input is checked before any work runs, which also catches the latent `"docker options"` reportType bug at the API boundary.
`dokku network:set --global bind-all-interfaces` (no value) wrote the literal string `false` to the property file instead of unsetting it, so external tooling could not distinguish "explicitly set to default" from "not set". Drop the empty-to-`false` coercion in `CommandSet` so the empty-value path reaches `PropertyDelete` like every other property, and remove the unconditional `TriggerPostCreate` write so new apps consult the global value before falling back to the `"false"` computed default.
Adds the missing raw/global/computed report keys for ten settable properties across the ps, builder, scheduler, proxy, openresty, nginx, and scheduler-k3s plugins so external tooling can verify drift through `:report --format json`. The scheduler-k3s `token` is masked as `*******` in default stdout output and only unmasked when the report is requested via `--format json` or when the flag is queried explicitly by name; the same explicit-query unmasking rule is extended to the existing traefik `dns-provider-<env_var>` keys.
The registry plugin accepted `registry:set <app> image-repo-template <value>` and persisted the per-app file, but only ever read the global value when rendering the image repository, so per-app overrides were silently dropped. The property now follows the same per-app to global to default fallback chain used by `push-on-release` and `server`, and `registry:report` exposes the raw per-app value alongside the existing global and computed flags. The `push-extra-tags` report flags were extended for the same reason: the runtime already honored per-app and global values, but the report only surfaced the raw per-app entry.
The longest `ps:report` flag is now `--ps-computed-stop-timeout-seconds` (34 chars), which pushes the per-plugin stdout column width past the 31-char floor used by `common.ReportSingleApp`. The `(report) report` test's literal-string assertion for `Deployed:` needs three extra spaces to match the new alignment.
The `ps` and `cron` plugins emitted `:report --format json` keys that
skipped the `<plugin>-` prefix or `-global-` infix used by every other
plugin, forcing external consumers to special-case those names. Rename
the affected report flags so they match the `<plugin>-global-<property>`
and `<plugin>-computed-<property>` convention introduced in 0.38.0, and
add the previously missing `computed-` siblings to both the per-app and
`--global` reports so the shape lines up with caddy, haproxy, and
traefik.
Reverts the earlier triggerRestart change. Forwarding the numeric tag forces
release_and_deploy to look up `dokku/<app>:N` (the bare local name), which
was never tagged when image-repo overrides the default. `fn-verify-app-image`
then retag-after-pulls and the underlying image ID ends up with two tags,
which prevents scheduler-docker-local's image retire path from removing the
old numeric tag (it uses unforced `docker image remove`).
When `image-repo` is overridden, the registry plugin pushes under that name and
the bare local `dokku/<app>:N` tag is never created. `fn-verify-app-image` now
queries the `deployed-app-image-repo` trigger and pulls `${REMOTE_REPO}${DEPLOYED_REPO}:${TAG}`,
retagging it as the local bare reference so subsequent inspects resolve. Also
aligns the `storage:ensure-directory` deprecation test assertion with the
standardized `Deprecated:` prefix.
Adds an app-aware `fn-verify-app-image` helper that pulls the deployed image
from the configured registry when it is missing locally, gated on
push-on-release being enabled, honoring per-app `registry:login` credentials.
Every in-tree `verify_image` caller migrates to the wrapper and the bare
helper is marked deprecated. `config:set` now forwards the deployed image tag
to `release-and-deploy` so the release path no longer pins recovery on the
unrecoverable `:latest` tag.
Three plugins still leaked the built-in default through `<plugin>-global-<property>` after the broader raw/computed split in #8640. External tooling reading `:report --format json` for drift detection could not distinguish "set to default" from "never set" - after a `:set --global` followed by an unset, the JSON still reported the default value. The `<plugin>-global-<property>` keys now hold the raw stored value (empty when nothing has been set) and `<plugin>-computed-<property>` keys hold the effective value with per-app, global, and built-in default fallback. Every global property now has a computed sibling for shape consistency, including those whose default is empty. Internal callers in `scheduler-k3s` that previously read the default-applied global helpers now consume new `getComputed*` helpers, keeping deploy-time behavior unchanged.
The `openresty:report --format json` payload emitted bare property names for the four global-only properties (`image`, `letsencrypt-email`, `letsencrypt-server`, `allowed-letsencrypt-domains-func-base64`), while `hsts` already used the `global-`/`computed-` convention adopted by the other proxy plugins. External tooling consuming the report had to maintain a per-property naming map within a single plugin. The global flag map now emits `--openresty-global-<X>` (raw stored value, empty when unset) and `--openresty-computed-<X>` (effective value with the built-in default applied) for those four properties plus `hsts`, and the per-app map gains the same pair for the global-only properties. The compose template and the two non-report callers consume the `computed-` accessors so the runtime behavior is unchanged.
nginx 1.25.1 deprecated the `listen ... http2` parameter in favor of a standalone `http2 on;` directive. Apps that ship a custom `nginx.conf.sigil` forked from an old default still hardcode `listen ... ssl http2;` and trigger `nginx: [warn]` lines from `nginx -t`. Surface a deprecation warning during `nginx_build_config` whenever the active template is app- or plugin-supplied and still contains the deprecated parameter, matching the existing warnings for `DOKKU_APP_LISTENERS`, `NGINX_SSL_PORT`, and `NGINX_PORT`, and point users at the `HTTP2_DIRECTIVE_SUPPORTED` variable already used by the default template.
The docker-local scheduler wrote each app.json cron command verbatim into the dokku user's crontab, where cron's `bash -c` interpreted any shell metacharacters in the command on the host as the dokku user rather than inside the container. The crontab line is now `dokku cron:run <app> <cron-id>`, and the command is resolved from app.json and exec'd inside the container at run time. Commands containing shell operators are also rejected when app.json is validated at deploy time.
The Traefik 26.0.0 chart registers CRDs under both `traefik.containo.us` and `traefik.io`, so a bare `kubectl get ingressroute` resolves against the legacy group where no resources exist and returns NotFound. Querying `ingressroutes.traefik.io` targets the group Dokku's chart actually writes to.
The `dokku.com/cron-id` label could exceed Kubernetes' 63-byte cap because the cron ID is `base36(appName === command === schedule)`, which expands roughly 1.5x per byte. The label is now keyed `dokku.com/cron-hash` and holds the `sha1` hex digest of the cron-id, a fixed 40-character value that always fits the cap. The same hex digest is mirrored into the `dokku.com/cron-hash` annotation, and the original base36 cron-id stays in the `dokku.com/cron-id` annotation that `cron:list` reads when surfacing user-facing IDs. Per-task lookups stay server-side via label selectors, so `cron:set --maintenance` and the forbid/replace concurrency checks on `dokku run --cron-id` keep working without any in-memory filtering.
Adds four Go template tests for ingress-route.yaml covering the imported-cert secret name, multiple domains per IngressRoute, the non-traefik ingress_class guard, and the https port_map dedup when an http port_map already covers the same container port. Also adds a bats integration test that deploys a TLS-enabled app under traefik and asserts both IngressRoute manifests and live http-to-https redirect plus https serving.
The canonical encode pipeline `echo "value" | base64` silently appends `\n` to the value, which round-trips through base64 and ends up stored in the env. Document that decoded values are stored byte-for-byte, recommend `printf '%s'` or `echo -n` for short inline values, and show `cat | base64 -w 0` for files. Adds a bats test asserting all four encoding patterns produce the expected stored bytes.
Closes#8647.
The asymmetric --logs-vector-global-image and --logs-vector-global-networks flags were renamed to --logs-global-vector-image and --logs-global-vector-networks in #8640 so they match the standard `--<plugin>-global-<property>` shape used by every other report. The bats description string at tests/unit/logs.bats:533 was still pinned to the old asymmetric phrasing, and there was no assertion that the deprecated JSON keys had actually been removed. Updates the test description to match its assertions and adds a regression guard that fails if logs-vector-global-image or logs-vector-global-networks ever reappear in the --global JSON report. Closes#8632.
The docker-options DOCKER_OPTIONS_<PHASE>.migrated sentinels and the storage data/storage-registry/migrations/<app> flag files are now stored as per-app properties (`migrated-<phase>` and `legacy-mounts-migrated` respectively), so they survive a property-store backup/restore and align with the existing `migrated-from-files`, `env-migrated`, and `nginx-conf-sigil-migrated` patterns. The new markers are written only when a non-empty legacy source was actually migrated, distinguishing apps that never had legacy state from apps that did and were drained. A one-cycle upgrade-cycle conversion drains any leftover filesystem sentinels into the new properties and removes the files; for docker-options the conversion deliberately runs before the global `migrated-from-files` short-circuit so installs upgrading from the previous release still get their `.migrated` files converted.
