Greptile review follow-ups on the paged end-user cache invalidation.
UserApiKeyCache keeps hashed token keys in a second in-memory partition, and
routes delete_cache / async_delete_cache there. It inherited the new batch
delete unchanged, so a budget cascade cleared the main partition and left the
key object sitting on its pre-reset spend. Override it the way
async_set_cache_pipeline already partitions its entries.
The spend counters and the management cache shared one exception handler, so a
Redis failure on the counters returned before the management cache was touched
at all. Each cache gets its own await and its own handler now.
A failed page read returned the same empty tuple that ends the walk normally,
so a truncated pass was reported as a complete one. The window is advanced by
then and no later tick comes back for the customers past that page, so the walk
now says it was cut short and the service log carries it.
Keep /projects list paging as a pushed history entry, validate the
project key table page size against its offered options, and clear the
key table params through the table-state setters instead of a copied
key list.
The projects list search lives in ?project_search= and its pagination now goes through useUrlTableState, keeping the page and page_size keys. The key table inside a project reads keys_search, keys_page and keys_page_size, resets to its first page on a new search, and no longer snaps a deep-linked page while the key fetch is failing. Closing a project drops its keys_ params so they do not leak into the next project
The organizations list now reads its search (org_search), org ID filter (filter_org_id), sort (sort_by, sort_order) and pagination (page, page_size) from the URL through useUrlTableState. The detail view tabs are controlled by ?org_tab=, and the Edit row action opens ?org=<id>&org_tab=settings in one history entry instead of passing an editOrg flag
A request for a configured model group that the router rejects before picking a deployment (all deployments in cooldown, no healthy deployment) never gets a custom_llm_provider in its logging kwargs. The spend log payload persisted an empty provider, the daily spend tables carried it through, and the Admin UI Usage page rendered those requests under unknown even though every model in the group has a provider
get_logging_payload now takes the proxy router and, when the logged provider is missing, infers it from the model group's deployments. It only attributes when every deployment in the group resolves to the same provider; mixed groups, unknown groups and a missing router leave the value empty as before. Explicitly logged providers keep precedence
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
The agent-runtime branch of /bedrock/{endpoint} (agents, knowledgebases, flows,
retrieveAndGenerate, rerank, generateQuery, optimize-prompt) forwarded every
caller header to AWS next to the SigV4 signature, so a LiteLLM key presented in
x-api-key or x-litellm-api-key reached bedrock-agent-runtime verbatim. Build the
upstream header set explicitly: drop LiteLLM credential headers by name and any
authenticated secret by value, keep the remaining caller headers, and let the
signed headers win on collisions.
It drives the real CLI for several seconds. The edge stamps every
upstream call with PYTEST_CURRENT_TEST, a process-global that names
whichever test the worker is in when the call arrives rather than the one
that made it, so a test that holds a worker that long collects other
tests' in-flight calls. Build 234's key report credits this test with 20
Bedrock and 7 Anthropic misses, and it makes no provider call at all.
Those misattributed calls take the wrong test id into the cache key and
write recordings under it, so the test was polluting the shared corpus it
exists to protect.
Deselected unless E2E_CLI_DETERMINISM is set, the same opt-in shape the
managed-files, prompt-caching and redis-chaos markers already use. The
attribution bug itself is older than this branch and is reported, not
fixed here.
A column toggle that localStorage refused was kept in memory and read ahead of storage, so a later save from another tab stayed hidden until this tab saved again. A storage event now drops the in-memory copy for its key, or all of them when another tab clears storage
usePersistedColumnVisibility kept a useState copy seeded once at mount, so a later tableId or defaults change showed the old table's columns and saved them under the new key. It now reads localStorage through useSyncExternalStore, keeping only writes that storage refused in memory, so the hook has no copy to go stale. Stored choices are layered over the defaults on every read, and changes saved in another tab show up.
Controlled client-mode DataTables no longer let TanStack reset the page index when rows change, since the owner of the pagination state decides the page. Once rows settle, a page past the end snaps back to the last page, matching server mode
Add useUrlTableState (search, sort, page, page size and filter_<column>
in the query string via one nuqs useQueryStates call, with keyPrefix and
urlKeys for routes that host two tables or need legacy key names) and
useUrlTab (validated ?tab= param with a role-aware fallback). Migrate
the Virtual Keys table onto useUrlTableState with a byte-identical URL
contract and bind the Playground tab strip to ?tab=.
DataTable gains controlled columnVisibility/onColumnVisibilityChange
plus usePersistedColumnVisibility (localStorage per table id), and an
isError prop that keeps the server page clamp from rewriting a
deep-linked ?page= after a failed fetch. Virtual Keys uses both.
The expired-session redirect in handleError now keeps the query string
and hash so the return URL captured on re-login restores the filtered
view instead of the bare list.
Delete useTabRouting and tabRoutes, the pathname tab router left over
from the reverted path-per-tab attempt (#34327, reverted in #34629);
tab persistence has to be a query param on the static export.
Interrupted Anthropic streams that die before message_delta were billed at
the message_start placeholder (any value above 1 was trusted) or at 0 when
the partial response was reasoning-only, because the token_counter fallback
only looked at visible text. Reset the placeholder whenever no finish_reason
or second usage event arrived, fold the already-counted reasoning tokens into
the fallback estimate, and drop the stale completion_tokens_details so cost
is computed from the recovered count
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
