Every attention row on every install (and all 38 rows on the public demo)
carried a warning 'Evidence incomplete; timing unavailable' / 'Evidence
timing unavailable' metadata line. Unknown freshness just means the evidence
source publishes no validity window - only availability-poller evidence does
- so it is not an operator-facing warning, and the 'timing unavailable' claim
was false with per-envelope observation times rendered beside it.
- Presentation: the scan row now surfaces an evidence phrase only for states
worth a glance (unavailable, incomplete, stale, or window-backed current);
complete evidence with unknown freshness stays quiet on the row and shows
a muted 'Evidence recorded' badge in the detail, mirroring the existing
unavailable-protection row rule.
- Alerts bridge: provider incidents without a backing availability check now
carry a first-class complete/confirmed envelope derived from the observed
incident payload instead of falling to the legacy partial shim, and each
sync merges the cycle's freshly observed envelopes into the active alert so
raise-time evidence does not age out while the condition is still observed.
Verified live on a scratch mock instance at 1280px and 744px: all 44+ mock
attention rows now project complete evidence and render without the noise
line; detail shows 'Evidence recorded' with per-observation timestamps.
The Plans & Billing navigation entry is suppressed for sessions without
commercial context, which hid the only activation-form entry point from
exactly the customers who need it: a fresh Pro install has no license yet,
so hideUpgrade stayed true until the key the customer could not find a
place to paste was pasted. The compiled Pro binary is only distributed
through the paid broker flow, so a Pro-edition session is commercial
context by construction — edition.IsPro() now sets commercialContext in
securityStatusPresentationPolicy.
Ordinary free self-hosted sessions run the community binary and keep the
opt-in default (the edition marker keys off the compiled binary, never
license state). Demo mode and white-label keep full suppression via
hideCommercial, which wins over edition-derived context. Complements
1b9bc9482 (setup-completion activation pointer); together a fresh Pro
install gets both the first-run pointer and a permanently visible
Plans & Billing entry. api-contracts, storage-recovery, and
agent-lifecycle contracts record the policy delta.
Clears the red main streak that began at 63a0adf9a:
- newTestConfigHandlers never stopped its alerts.Manager, leaking a
periodicSaveAlerts goroutine that logs through the global zerolog
logger and races with the authorization-refusal contract test's
logger swap under -race
- InfrastructureAgentDoctorPage used a raw select, tripping the
settings native-select guardrail; converted to the shared labelled
FormSelect primitive
- the metric-title residual pin still expected the removed -1 sentinel
copy; aligned with the Off-control wording from 16179dd0a
Contract-Neutral: test-only CI red fix: stops alerts.Manager goroutine leak in api test helper, converts agent doctor raw select to the canonical FormSelect primitive the frontend-primitives contract already mandates, updates a stale test pin; no public contract delta
Route config-owned install commands and setup-script artifacts through the canonical trusted request-origin resolver. Preserve configured URL precedence and add endpoint-level adversarial coverage for token-bearing commands.
Agent install and update commands advertised http://<LAN-IP>:7655 on
installs where the operator browses Pulse over an https FQDN, because
boot-time detection pre-fills config.PublicURL with an IP guess and
resolvePublicURL treated that guess as configuration. Track
auto-detected values (boot probe or first-request capture) separately:
explicit config from the env var or the settings UI still wins, but an
auto-detected guess now yields to the origin of the admin request that
is asking, so copied commands match how the instance is actually
reached.
Also stop boot detection from clobbering a settings-persisted public
URL on restart, which silently reverted the operator's saved value
whenever PULSE_PUBLIC_URL was unset.
Refs #1692
Contract-Neutral: behavioral fix to advertised-URL derivation: no API shape, field, or endpoint changes; PublicURLAutoDetected is internal (json:-)
755a88878 gated Settings -> Infrastructure on infrastructureRead but left the
System group untouched. Network, Pulse server updates, and Recovery are the
same shape: pure instance administration - the public URL and CORS boundaries,
the server update channel, backup polling and configuration export/import -
behind routes that are all RequireAdmin + settings:read. A session without it
was still offered all three, could deep-link straight at them, and landed on
panels it can never populate.
Serve systemSettingsRead alongside infrastructureRead and gate the three tabs
on it. requiredCapability blocks the route as well as the nav item, so a deep
link now falls through to the capability-aware fallback that 755a88878 added,
landing on the first tab the session can actually reach.
The new field is a sibling of infrastructureRead, not a reuse of it. They
share the canAccessAdminSurface(config.ScopeSettingsRead) derivation today, but
infrastructureRead is named and documented for the data sources behind one
page. Borrowing it to gate a different page is the same defect as re-deriving
a capability: the two would drift the first time one surface's gate changed,
and the api-contracts rule on published capabilities exists to prevent exactly
that. Each field names the surface it gates and is pinned separately.
These hide rather than render an inline gate, matching Infrastructure rather
than the paid-feature items in the same catalog. A free install seeing a paid
tab can act on it by upgrading, so hiding it would cost discoverability. A
non-admin cannot grant themselves admin, so the tab is only noise.
system-general stays ungated on purpose: theme, language, and unit preferences
there are user-scoped, so gating it would take personal settings away from
every non-admin, and the panel is not empty for them.
Verified against real admin and non-admin sessions rather than a stubbed
capability - one binary built from this worktree behind two header-injecting
proxies differing only in the role header. The viewer sees none of the three
tabs and a deep link to /settings/system-network lands on
/settings/monitoring/availability; the admin keeps all three and the same deep
link resolves to the Network panel. Confirmed at 1280x720 and 375x812.
A refusal is the access control working, not a fault, but every one of them
warned twice: once in RequireAuth/RequireAdmin/RequirePermission, and again in
the middleware, which warned on every 4xx unconditionally. A correctly
configured instance therefore could not produce a quiet log, and #1601's rc.9
reporter read that stream as an RBAC regression.
Gating clients one at a time does not fix this. The previous commit stopped six
endpoints being polled by a non-admin UI, and /api/updates/status,
/api/updates/plan and /api/availability-targets still warned, as would every
caller nobody has audited yet.
Refusals now route through logAuthDenial, which records them at debug and counts
them per caller. Attribution prefers the authenticated username so a principal
stays tracked across rotating addresses, falling back to the client IP. Crossing
20 refusals in a minute emits exactly one warn for that window, which is the
shape that separates probing from a UI mounting a surface its session cannot
read; a closed window re-arms it. The tracked set is bounded with oldest-window
eviction so spoofed forwarded-for values cannot grow it. The middleware now
warns only on 5xx.
Enforcement is untouched: every route returns the same status to the same
callers, and the contract test pins that pairing so a future attempt to quiet
the log by relaxing enforcement fails rather than passes. Verified live on a
proxy-auth instance - /api/connections, /api/updates/status and
/api/system/settings still 403 for a viewer and 200 for an admin; 19 refusals
produce no warn, the 20th produces one, and 30 more produce none; an idle
non-admin browser session logged zero warn lines across 90 seconds.
Settings -> Infrastructure reads /api/connections, /api/config/nodes,
/api/system/settings, /api/truenas/connections and /api/vmware/connections on
mount and then polls /api/connections every 15s and /api/discover every 30s.
Every one of those is RequireAdmin, so an authenticated non-admin rendered a
page where nothing loaded while each poll reprinted "Non-admin user attempted
to access admin endpoint" at warn level. Measured on a proxy-auth viewer
session: 6 denials/minute from this page alone on an idle tab.
Serve infrastructureRead alongside the other settings capabilities, derived
from the same canAccessAdminSurface(settings:read) expression the routes
enforce, and gate the nav item on it. Two follow-on fixes were needed because
the page is not the only mount point:
- DEFAULT_SETTINGS_TAB is infrastructure-systems, so the blocked-route
fallback pointed straight back at the tab it had just refused. It now falls
back to the first tab the session can actually reach.
- Settings.tsx constructs useInfrastructureSettingsState for every settings
tab, so the discovery poller and the TrueNAS/VMware mount fetches ran no
matter which tab was open. They now wait on the same capability. The
TrueNAS/VMware loads moved from onMount to an effect so admins still load
once the capability resolves, rather than sampling it before it exists.
Verified against a local instance behind a header-injecting proxy-auth shim:
viewer goes from 6 infrastructure denials/minute to 0, admin keeps both
pollers armed (/api/discover x4 and /api/connections x3 over 152s) with 0
denials. Re-checked at 1280x800 and 375x812 for both roles.
Stripe checkouts could not be told apart by origin: every session was
stamped checkout_origin pulserelay_landing, so an upgrade started from an
RBAC gate looked identical to one started from the website. Gate CTAs now
stamp a closed-vocabulary source token (gate-<feature>, estate-card, or the
plans-page default) onto the owned billing plan route; the plan page threads
it into the purchase-start handoff, and the license server persists it on the
checkout intent and stamps Stripe metadata checkout_source.
Attribution is authenticated-session-only by construction. The public
/pricing route and the public pricing URL never carry it, and
getSelfHostedPurchaseStartUrl scrubs the parameter from forwarded query
strings so a crafted website link cannot claim in-app origin. Server-side,
source is validated against the same kebab vocabulary, skip-listed from the
Pulse Account portal redirect exactly as feature already is, and travels
only inside the handoff body. It is request-scoped and persists nothing on
the install; the cancel return echoes it so a retry keeps its origin.
The handoff field is omitempty because the license server decodes strictly,
so source-less installs stay compatible with a server either side of the
field being added. That ordering is recorded in pulse-pro OPERATIONS.md.
The 2026-08-07 telemetry read showed installs at or above 5 PVE nodes, 10
Docker hosts, or 3 VMware hosts convert to paid at ~8x the rate of smaller
estates. The ping now carries that classification as a server-derived
boolean so receiver-side cohort queries keep a stable column even if the
thresholds move later; it is derived in the pkg/server snapshot closure
from the same AggregateInstallSnapshotCounts values the payload already
sends, so no new information leaves the install.
The thresholds move to internal/monitoring/business_estate.go as the
single definition; the session-capability surface behind the in-product
business-estate card delegates to it, and dropping the now-unneeded direct
pkg/licensing import there restores TestPkgLicensingImportBoundary, which
f0e2243b4 had left red. All three payload surfaces (Ping struct, private
receiver, TelemetryPingPreview) move together per
check_telemetry_schema_parity.py, and both PRIVACY.md copies document the
field.
Verified live on an isolated worktree backend with 6 mock PVE nodes: the
Settings telemetry preview renders schema_version 8 with business_estate
true, and /api/security/status still reports
sessionCapabilities.businessEstate true through the delegated thresholds.
Free self-hosted sessions now see reactive commercial surfaces: paid-feature
nav items stay visible with panel-owned inline gates (Relay precedent extended
to reporting, roles, users, audit log, audit webhooks), gate CTAs render, and
Plans & Billing is discoverable and carries an MSP/provider section with the
free 60-day evaluation. Adds a one-shot business-estate card for authenticated
free sessions whose monitored estate crosses business-scale thresholds
(>=5 PVE nodes, >=10 Docker hosts, or >=3 VMware hosts); the flag rides
sessionCapabilities only, never the pre-auth presentation policy, so estate
size cannot leak to anonymous visitors. Demo mode and white-label runtimes
(including MSP tenant containers) keep full commercial suppression; the
multi_tenant organization nav items deliberately stay hidden.
Decision record supersedes the 2026-04-25 opt-in record:
docs/release-control/v6/internal/records/self-hosted-commercial-surfaces-revision-2026-08-07.md
RA5 summary, contract pins, nav integration pins, and the RA5 browser spec
are updated to the revised invariant in the same commit.
The server now echoes its version on unified-agent report acks, and the
agent nudges its auto-updater the moment an ack carries a newer version.
After a server upgrade, agents converge within one report interval
instead of waiting out the hourly update check, so the "older Pulse
agent" notice self-resolves in seconds once upgraded agents report in.
The hourly loop stays as the retry and backstop path. Nudges dedupe per
server version, refuse downgrades, skip disabled and development-mode
updaters, and never fire from observer destination acks — only the
authoritative server may steer an agent's updater, and a nudged check
re-validates against the server and runs the existing checksum and
self-test pipeline before swapping binaries. Agents deployed before
this change still converge on their old hourly cadence once; every
upgrade after that lands within a report cycle.
Contract deltas recorded in agent-lifecycle and api-contracts, with
boundary notes in security-privacy (no update authority in the echo),
performance-and-scalability (no steady-state work), and
storage-recovery (nothing persisted).
A pulse-agent installed on a Proxmox node merges into one canonical
resource whose primary ID is node:{sourceID}, so the agent:{hostID}
reference the alert evaluator uses (hostResourceID) matched nothing in
the registry. Per-resource grace overrides and operator-state lookups
then silently resolved to factory policy: alerts fired with no grace
while the policy preview, which receives the canonical ID directly,
showed the override working. Index the prefixed agent ref as a
canonical identity alias so runtime resolution lands on the same
resource the override was saved against.
Refs #1497
Contract-Neutral: behavioral fix#1497: index agent: ref as canonical alias, no public contract shape change
restoreProviderMSPArchiveFile copied each tar entry with an unbounded
io.Copy, so the only limit on what a restore wrote to disk was the size
the archive declared for itself. A gzip bomb, a PAX sparse entry claiming
a huge logical size, or a corrupt stream could fill the target volume.
Bound extraction the way readProviderMSPBackupManifestBytes already
bounds the manifest: a per-entry cap and a cumulative cap across the
whole restore, both enforced against the bytes actually copied rather
than the declared header size. That size comes from the archive, so it
is only good for an early reject, never as the bound. An entry that
overruns fails the restore and its partial file is removed, rather than
being silently truncated into a file that looks complete.
A restore that failed partway had already deleted whatever it replaced,
leaving a half-populated control plane that looks bootable and forcing
the retry to use replace. Roll the partial restore back to an empty
target instead, and say so in the error.
Follows the zip-slip fix in the same function (CodeQL alert 314).
The #1601 follow-on: per-container alert overrides were keyed by Docker
container ID (docker:{host}/{containerID}), which changes on every
recreate, so each image update silently re-armed alerts the user had
switched off and left a dead entry behind in alerts.json — the unbounded
growth that pushed the reporter's config past the old 64KB body cap
(raised in 38434a513). The v6 thresholds UI additionally wrote keys from
the unified hash id (docker:{host}/app-container-{16hex}), which the
evaluator never read at all.
Overrides now key on stable identity, docker:{host}/{containerName}:
- The evaluator resolves the name key first and falls back to the legacy
container-ID key so pre-migration entries keep working
(evaluateDockerContainer, checkDockerContainerState, the
container-update resolver, and reevaluateActiveAlertsLocked).
- MigrateDockerContainerOverrideKeys runs in the monitor sync next to
MigrateCanonicalOverrideKeys, driven by the unified resource snapshot:
it re-homes live legacy-ID and unified-hash keys onto the name key and
prunes orphaned ID-shaped entries, ending the per-update orphan
accumulation. Name-keyed entries for absent containers are kept so a
recreate under the same name still honours them.
- The UI candidate chain (single implementation in alertOverridesModel)
now leads with docker:{host}/{name} and trails the container-ID, short
ID, unified-hash and slash-tail forms, so rows bind pre-existing
overrides of every historical shape and the next save re-homes them.
Rows carry overrideStorageId/overrideIdCandidates so toggle,
connectivity, offline-state, edit and remove all write the stable key.
- The ignored-containers card copy now documents the wildcard forms
(runner-*, *-dev, *staging*) shipped in b5fa6a9af, under the title
"Ignored container patterns".
Contract deltas: alerts, frontend-primitives, monitoring, and
unified-resources now pin the name-keyed override identity, the single
frontend candidate-chain owner, the sync-cadence migration, and the
resource-facet-backed table identity respectively.
Verified live against a mock instance: a UI toggle persists
docker:{host}/loki and binds back after reload, and seeded
legacy/hash/orphan keys converge to name keys on disk within two sync
ticks. go test ./internal/alerts/... ./internal/monitoring/... green;
recreate survival pinned in
TestDockerContainerOverrideSurvivesContainerRecreate.
Refs #1601
The PVE and PBS branches of handleCanonicalAutoRegister each carried a
verbatim copy of the identity-matching ladder — exact host, then TLS
fingerprint conflict rejection, resolved host identity, then DHCP
continuity by node name plus token. golangci-lint's dupl flagged the pair.
Extract one matcher, findCanonicalAutoRegisterMatch, over a
canonicalAutoRegisterCandidate identity view that both instance types
project onto. Behavior is unchanged: the same ladder in the same order,
the same log messages and fields, and the same (index, preserveHost)
result, with the previous "break with preserveHost still false" paths
becoming explicit returns.
golangci-lint run ./... green; full go test ./internal/api/ green.
Contract-Neutral: dupl-only refactor of handleCanonicalAutoRegister: PVE/PBS identity-matching ladder hoisted into one helper, zero public-contract or behavioral delta
The vSphere adapter filled `Resource.Tags` with six fixed strings on every
resource — `vmware`, `vsphere`, `<kind>`, `source:vcenter`,
`connection:<name>`, `power:<state>` — and never read vCenter's own tag and
category system. Every VM in an estate returned a byte-identical set, so the
workload Tags column rendered the same dots on every row and filtering on any
of them selected everything. Commit 6b78feba8 default-hid the column and said
in as many words that the hide was a stopgap awaiting this fix.
`internal/vmware/client_tags.go` reads the CIS tagging service. That is a
different endpoint family from the `/api/vcenter/...` inventory reads, but the
same vSphere Automation API, so it reuses the caller's `/api/session` token
rather than opening and managing a second session. Associations come from one
batched `list-attached-tags-on-objects` POST per bounded object batch, never a
per-object request; tag and category names resolve through a client-scoped
catalog with a 10-minute TTL, so a steady-state refresh of a tagged estate
costs only the association reads while a rename still converges without a
restart. A vCenter without the tagging service, or an account without the tag
read privilege, degrades into a `tags` stage enrichment issue and leaves the
inventory untagged; it never fails the refresh.
The provenance strings stay. `Resource.Tags` is the only keyword set
`resourceSearchMatch.ts`, the `?tags=` resources filter, and saved
report-schedule tag filters read — `collectSearchCandidates` gathers no
`technology`, `type`, or `platformScopes` candidate — so dropping "vmware" or
"vsphere" would silently stop matching searches and saved filters that depend
on them. Real vCenter labels are appended to that set, never substituted for
it.
Because that flat set is deliberately mixed, it is the wrong source for a
per-row Tags cell. Real tags therefore also land on a canonical `VMware.Tags`
facet that carries vCenter's category alongside each name, and
`useWorkloads.ts` maps `WorkloadGuest.tags` from that facet for any resource
carrying VMware metadata — including the empty case, so a vSphere VM nobody
tagged renders an empty cell instead of falling back to the provenance dots.
vCenter tag names are unique only inside their category, so the flat label is
`category:name`: two categories may each hold a "Production".
With the column carrying per-row meaning again, `tags` leaves
VMWARE_WORKLOAD_DEFAULT_HIDDEN_COLUMN_IDS and the `defaultHiddenMigrationIds`
retirement list, and the state-model test that pinned the stopgap now pins its
absence. No un-hide migration ships alongside it: 6b78feba8 is on main but no tag
contains it, so the stopgap never shipped and no install carries the
auto-hidden preference. That holds only while the two stay together — the
migration writes the hide into each user's saved preference on first load,
so an rc cut from main carrying the stopgap without this commit would make
an explicit un-hide path necessary.
Mock fixtures carry uneven tag coverage — several categories on some objects,
one on others, none on the rest — because a uniform fixture set would hide
exactly the defect this data exists to catch.
Verified against a mock estate built from this branch: `/api/resources`
returns provenance plus real labels on the flat set and only real labels on
`vmware.tags`; the Tags column renders 2-4 dots per tagged VM and none for
untagged ones; a dot's tooltip reads `Backup:Nightly`, and clicking it
searches `tags:Backup:Nightly` and narrows 18 VMs to the 3 that carry it.
Contract deltas: performance-and-scalability.md Extension Point 17 replaces
the stopgap paragraph with the two-surface tag contract and the bounded
tag-read budget; unified-resources.md states the keyword-union vs facet split
and that a present-but-empty facet means "no operator tags" rather than a
fallback; storage-recovery.md extends its VMware descriptive-only boundary to
`vmware.tags`, because vCenter tag vocabularies read like protection policy
(`Backup:Nightly`) and a label the operator wrote must never satisfy a
coverage or compliance verdict that recovery-owned evidence should decide.
Resolves every open code scanning alert on the repository. Dependabot and
secret scanning were already clear.
SMART temperature truncation (alerts 312, 313). parseRawValue returns a
64-bit raw attribute value, but DiskSMART.Temperature is an int, which is
32 bits wide on the 386 and arm release builds Pulse ships. The range check
ran after the narrowing conversion, so a raw value of 4294967316 truncated
to 20 and was published as a plausible 20 degree reading.
validSMARTTemperature64 now gates the conversion.
Provider MSP restore archive names (alert 314). cleanProviderMSPArchiveName
rejected a leading "../" but not a bare "..", which path.Clean produces from
entries such as ".." and "a/../..". pathIsInside caught the escape
downstream, so this was not exploitable, but the sanitizer now rejects it
outright instead of depending on a second gate.
TrueNAS device paths (alert 315). vdev.Device is supplied by the appliance,
concatenated into a path and published verbatim on ZFSDevice.Path, so values
like "//evil.example.com/share" and "/\evil.example.com" passed straight
through. devicePath now drops traversal segments and backslashes and
collapses a leading double slash. The alert's open-redirect framing does not
apply here, there is no redirect sink on this path, but the value is
untrusted input rendered as a path and is worth normalising.
Patrol readiness cache key (alert 311). The key is persisted to
ai_patrol_model_readiness.json and embedded an unkeyed SHA-256 of the Ollama
Basic Auth username and password. That password is chosen by a human, so
anyone holding the evidence file could recover it offline at two SHA-256
operations per guess. The fingerprint is now HMAC-SHA256 keyed with a
32-byte per-install salt stored beside the evidence at mode 600. Credential
rotation still invalidates the cache and the key still survives a restart.
Each fix carries a regression test confirmed to fail against the previous
implementation.
monitoring.md carries the one warranted contract refinement. It already
required SMART temperature selection to accept only plausible readings, and
that rule now states the width at which plausibility is decided.
Contract-Neutral: CodeQL security fixes with no public-contract delta and no payload change. monitoring.md carries the one warranted refinement (SMART plausibility decided at 64-bit width). Residual demands are inapplicable: ai-runtime readiness prose documents interruption semantics, not cache-key derivation, and the credential-invalidation contract is unchanged; cloud-paid and deployment-installability contracts never name archive-entry sanitisation; agent-lifecycle owns smartctl.go but its SMART temperature prose lives in the staged monitoring.md.
Bare entries keep their historical prefix semantics. Entries may now
also use the wildcard forms already established by the PBS datastore
excluder, so *-dev matches a suffix, *staging* a substring, and
runner-* an explicit prefix. Wildcard-only entries are skipped, that
job belongs to DisableAllDockerContainers.
Suffix matching is what the reporter on #1601 was approximating with
hundreds of per-container disable toggles, which is also what pushed
their config past the old request body cap.
Refs #1601
Contract-Neutral: extend docker ignored-prefix matching with wildcard forms; no schema or payload changes
Alert config grows one Overrides entry per toggled resource, so the
64KB cap on PUT /api/alerts/config rejected saves from instances with
a few hundred disabled containers with 'http: request body too large'.
The 32KB cap on bulk acknowledge/clear failed ack-all during large
alert floods, the exact situation it exists for. Intent policies carry
per-resource rules with the same scaling shape.
All four now share a 1MB bound, which still caps memory per request
but no longer rejects legitimate fleet-sized payloads. Endpoint tests
pin a >64KB config save and a >32KB bulk ack at 200.
Refs #1601
Contract-Neutral: raise alert config and bulk ack request body caps; no payload shape or field changes
The 25 July GLM swarm generated these and they were never collected, unlike
the 0712 through 0724 batches already in main. They cover apiClient retry and
abort handling, the prompt-secret model boundary sanitizer, AI tool
normalisation, Patrol handoff, threshold table state, the audit log panel, the
connections ledger, licence and resource-badge presentation, and the AI
intelligence store.
Verified against current main before harvesting rather than trusting their
age: 36 Go tests and 197 frontend tests pass, go vet is clean, and both
batches were re-run after formatting.
rcourtman