The invite-dialog test only read data-orientation, so a regression inside
the shared field variants could restore the full-width bar and still pass.
Assert the classes that carry the layout, cover the two SSO call sites the
fix also changed, and pin the vertical/horizontal contract on the primitive.
bg-none only clears background-image, so the buttons fell back to the
browser's default button background instead of the transparent one the
inline style had.
key_edit_view opened a select and then clicked the option it found by
title text or by raw text. Both queries match the moment the option
enters the DOM, which is one render before the popup finishes entering.
Until then the positioner still carries an inline pointer-events: none,
and user-event refuses to click through it.
That is a race, and a fast machine loses it. Five of the file's 84 tests
failed on every local run while CI stayed green, which is the worst shape
for a test to have: it is only ever red on the machine of whoever is
trying to change the code.
tests/test-utils.tsx already ships chooseSelectOption for exactly this.
It finds the option by role and waits for the positioner to release
pointer events before clicking. The five call sites now use it, and the
helper takes the direct user-event API as well as a setup() instance so
callers do not have to restructure to use it.
Five consecutive full-file runs pass where every previous run failed.
Also finishes this file's screen queries, which brings
prefer-screen-queries to its target of 18.
The page painted every surface, border and text color inline with a
fixed light palette (#202124, #5f6368, #dadce0, #f8f9fa, #fff), so in
dark mode it drew dark text on hardcoded white cards.
Move the whole component to the foreground/muted/border/card/info
tokens, which already resolve for both themes.
The tool cards hardcoded light colors as inline styles (#fff, #fafafa,
#f0f0f0, #f6ffed), so in dark mode the theme's light foreground text
landed on a white card and became unreadable.
Swap the inline hex for the existing card/muted/border/success tokens,
which already carry both light and dark values.
Two changes, both about finding elements the way a user finds them.
Twenty-six test files destructured queries off render and called them
bare. Those queries are scoped to the render container, so they quietly
miss anything portalled into the body, and they read as if they were
free functions. They now go through screen.
ChatMessageBubble and the key info panel derived elements by walking
closest/parentElement/firstElementChild and then asserted on the classes
they found. A wrapper element anywhere in between broke them. The bubble
surface, the avatar and the budget reset value now publish a test id, so
the assertions survive markup changes and still fail when the styling
they check actually regresses.
Budgets drop with the counts: prefer-screen-queries 221 to 21,
no-node-access 723 to 716.
The 21 remaining prefer-screen-queries are not all fixable: 18 of them
are within(dialog) results in MCPToolsetsTab, which the rule cannot tell
apart from a render result. Target is 18, not 0.
The shared DataTable test reached for elements by CSS selector and by
walking parentElement chains, then asserted on Tailwind class strings. It
had no role queries at all, so a wrapper div anywhere in the render tree
broke it while changing nothing a user sees.
Columns, rows and headers are now found the way a user finds them: by
role and by the text on screen. The compact skeleton row is compared
against the loaded row's height rather than a hard-coded h-8, so renaming
the class no longer breaks the test but shrinking the row still does.
The fillHeight and maxBodyHeight cases stay class assertions. jsdom has
no layout engine, so there is nothing behavioural to assert there. What
they no longer do is derive their elements from incidental nesting: the
three layout wrappers and the header now publish a stable test id, which
is also why the resizer's write-only data-resizer attribute became one.
Budgets drop with the counts: no-container 150 to 133, no-node-access 760
to 723.
Turn on testing-library/no-node-access, no-container and
prefer-screen-queries as warnings and baseline them in eslint-budgets.json
so the counts can only go down.
These three rules catch tests that assert on DOM structure rather than on
what a user can observe: reaching through parentElement chains, querying
the container by CSS selector, and destructuring queries off render
instead of going through screen. Those assertions break on refactors that
change nothing a user sees, and stay green when the behaviour underneath
is broken.
Baselines are the current counts, so nothing fails today.
Classification timing and session affinity are the same operator question,
so Advanced: Classification Method now carries a single "How often to
classify" radio: every request, every new user message, or once per session.
The session choice writes session_affinity and stays disabled on custom tier
sets, where the backend rejects it. Advanced: Affinity keeps the deployment
switch alone.
The serializer always writes classification_mode, matching session_affinity
on the line below it, so an explicitly stored every_request survives an
untouched save instead of being dropped back to the backend default.
max_budget and reset_at already live on the matching budget_limits entry, so
repeating them (as budget_limit and reset_at) only invited confusion about which
copy is authoritative.
budget_limits now comes back exactly as stored on /key/info and /v2/key/info.
The per-window usage moves to a sibling budget_limits_usage field keyed by
budget_duration (current_spend, budget_limit, reset_at), mirroring
model_max_budget_usage, so the stored shape that /key/update accepts never
carries a computed field.
* feat(ui): auto-router controls for context-window escalation
Adds an Advanced: Context Window Escalation section to the auto-router
form, both create and edit arms, with the toggle for
enable_context_window_escalation and a clamped decimal input for
context_window_escalation_buffer. An untouched control keeps both keys
out of the payload so the router tracks the backend defaults; an
explicit opt-out (false) survives the edit round-trip through the
managed-keys projection and the hydrator, and preset prefill maps both
keys straight through so a preset cannot silently drop them
Resolves LIT-6601
* fix(ui): clearing the context-window buffer removes it from the payload
Both review bots converged on the same defect: an emptied buffer field
early-returned in commitBuffer, the draft was discarded on blur, and the
stale number reappeared and stayed in the saved config, contradicting
the copy that an empty field tracks the backend default. An empty commit
now removes the key, which the managed-keys projection propagates as a
real deletion on edit. Also trims the narrative comments the review
flagged as restating behavior
* fix(ui): keep litellm_credential_name from LiteLLM Params JSON when no credential is selected
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* test(ui): drop null litellm_credential_name from AddModelPanel payload fixture
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
---------
Co-authored-by: yassin <yassin@berri.ai>
Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Shadow eval jobs previously targeted only virtual keys, so deployments on
pure JWT auth (which present no key at all) could never sample their
traffic. Jobs now carry a typed (target_type, target_id) pair covering
keys, teams, and users; sampling matches the identity every request
resolves to at auth time, so team and user jobs cover JWT traffic with
no client changes.
Resolves LIT-6578
* feat(complexity_router): escalate oversized prompts to a tier that fits before dispatch
The classifier scores complexity and never prompt size, so a long agentic
session whose newest ask is trivial classifies SIMPLE onto a small-window
tier and the provider rejects it with a context-window 400 that nothing
retries. The gate runs after classification on every decision path
(classify tail and session-affinity pin), estimates prompt tokens
including the out-of-band carriers (top-level system, tools,
instructions), and when the decided tier provably cannot hold the prompt
moves the request to the lowest configured tier with a model whose
declared window fits, restricting the pick to fitting models when the
decided tier can keep it. Models with no resolvable window are never
escalated away from or onto, escalated decisions are never written as
session pins, and the decision records context_escalated plus the
original tier in spend logs.
Resolves LIT-6503
* fix(complexity_router): judge groups by smallest window, bound skips by bytes, filter adaptive picks
Review-round rework, one mechanism per finding. A group is judged by its
smallest resolvable deployment window, since the core router picks within
a group with no fit check. The counting skip is gated on UTF-8 byte
length, which BPE token counts can never exceed, so token-dense scripts
cannot slip past it; only a real tokenizer count ever moves a request and
a failed count leaves the placement alone. The fit facts now filter every
adaptive phase including cold start and the tier fallbacks. Window
questions adopt the declared provider and never resolve authenticating
providers, and a router instance without get_model_list degrades the gate
to a no-op. Tests rebuilt on real Router instances resolving deployment
model_info end to end, plus a full-path test through
async_get_available_deployment
The UI e2e suite had a class of assertions that pin how the dashboard is
built rather than what it does, so an ordinary refactor turns them red
without any user-visible change.
Geometry. The auto-router template select had two tests made of pixel
arithmetic plus a data-side="bottom" check, which is Base UI's own
positioner signal. The regression they guard (#38554) is a popup opening
on top of the control that spawned it, so both cases collapse to one
invariant: the options never cover the trigger. It now runs at both
viewport heights and reads the popup as role=listbox. The models header
test compared the tabs and refresh centers within 2px, which a padding
change flips; it now asserts the two share a row.
Structure. The logs drawer test walked xpath=../../.. from a text node
and read collapsed state off chevron icon classes. SectionHeader now
renders a real disclosure button with aria-expanded, and its two copy
buttons carry distinct names instead of both being "Copy". Sidebar group
toggles expose aria-expanded too, so the migration spec can ask for a
collapsed group by state rather than by nesting depth.
Positional lookups. keyRow.locator("button").first(), row.locator("td")
.first() and getByTestId(grid).locator("div").first() all named a
position where they meant an action; they now name the control. Table
scoping moves from "table tbody" to role=row.
Timing. Nine waitForTimeout calls are gone. Every assertion that followed
them already retried to its own timeout, so the sleeps only slowed the
run down.
Both files under tests/users/ were wrapped in test.skip("...", () => {}),
which registers one skipped test and never runs the body, so the four
tests inside had never executed and were written against a UI that has
since changed (the search placeholder is "Search by email…", the ID
filters moved into a drawer, pagination is labelled "Go to previous
page"). Rewritten against the current surface: the suite goes from 104
collected tests to 107.
Left in place deliberately: the chip and dialog-footer data-slot
selectors, because the accessible names they work around live in
components/ui/, which is shadcn CLI-managed and not hand-edited.
* feat(spend_tracking): persist router metadata in spend logs for internal router models
* test(spend_tracking): expect router_metadata key in exact-payload tests, type the routed-kwargs helper
* test(newrelic): cover static default_team_settings per-team routing
The dynamic POST /team/{team_id}/callback path for New Relic is tested, but
the static default_team_settings twin had no regression coverage. Add a test
that drives default_team_settings -> add_team_based_callbacks_from_config and
asserts the resolved trusted vars dispatch to BOTH the per-team metrics logger
(cost/usage) and the trace logger (LLM/agent spans), so a config-file customer
gets the same per-team routing as the API customer.
Also correct the /team/callback docstring: callback_name is a str validated
against the credential-capable callbacks, not a fixed langfuse/langsmith/gcs
Literal, and document the newrelic_api_key / newrelic_region vars.
* chore(ui): sync schema.d.ts with the /team/callback docstring
Regenerate the dashboard OpenAPI types for the add_team_callbacks description
change: callback_name is a validated str (not a langfuse/langsmith/gcs
Literal) and the newrelic_api_key / newrelic_region vars are documented.
* docs(newrelic): note LITELLM_OTEL_V2 prerequisite, trim test comments
Address review: team-scoped New Relic config is rejected with a 400 unless the
proxy runs with LITELLM_OTEL_V2=true, so document that in the /team/callback
endpoint and sync schema.d.ts. Drop the narrative setup comments in the new
test per the repo comment convention; the test name and docstring already say why.
* fix(proxy): deliver budget alerts on webhook-only alerting and accept ALERTING_WEBHOOK_URL
ProxyLogging.budget_alerts forwarded to the alerting pipeline only when
'slack' was in general_settings.alerting, so alerting: ['webhook'] plus
WEBHOOK_URL silently never delivered a budget alert (the config
/health/services?service=webhook exists to test). Forward when 'webhook'
is present too; SlackAlerting.send_alert already fans out per channel.
Also accept a provider-neutral ALERTING_WEBHOOK_URL env fallback for the
Slack-format channel (any Slack-compatible receiver works), mark it as a
sensitive var, and de-brand the admin UI alerting copy.
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* fix(ui): format settings.tsx with prettier
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* chore(ui): regenerate schema.d.ts for updated alerting description
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* ci: retrigger checks after ALERTING_WEBHOOK_URL docs merged
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
---------
Co-authored-by: yassin <yassin@berri.ai>
Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
The route-level RBAC in litellm/proxy/auth/route_checks.py 403s
/model/new, /model/update, and /model/delete for proxy_admin_viewer on
the session role alone, before ModelManagementAuthChecks' team-admin
carve-out can run. A view-only session therefore gets no model write
affordance, team admin or not.
The Add Model page's connection test dialog rendered a literal ", ]" beside the Close button, left over from converting an array of footer buttons into JSX children.
An Admin Viewer and a group whose name contains a slash both reached an
enabled Set budget action that could only ever come back 403 or 404. Gate
the row actions on proxy admin and on the name being addressable, with the
reason in the tooltip.
