A temporary-only member update no longer clones the team default budget into the private row. The row stores just the temp pair and auth, spend admission and reservation add the active increase to the current shared default, so a later lowering of the default reaches members with an active grant
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Adds temp_budget_increase and temp_budget_expiry to the team member edit form with pair validation,
seeds stored values into edit mode, sends both through /team/member_update, and adds cached-key auth
and reservation regression tests for active and expired increases
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
SpendLogsMetadata gains a typed azure_spillover key so a request Azure
served off pay-as-you-go capacity is visible in spend tracking, stamped
from the provider response headers or the processed llm_provider- headers
on the standard logging payload. The header parsing moves into a shared
azure_spillover() helper that is_spilled_over_ptu_request() now wraps.
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
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 reconcile now records the database clock of the scan behind the last complete
run and, on the next run, rewrites every closed day with per-key rows updated since
then, however old the day is. Replaying only the marker day and the one before it
missed a delayed flush or retry that landed on an older date, and reads through the
marker come from the global table alone, so that spend was never counted.
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
A team can now carry a per-model budget map that every key on the team
inherits. A key's own model_max_budget entry for the same model takes
precedence, so it is gated on and billed to the key alone.
Backend: NewTeamRequest/UpdateTeamRequest accept model_max_budget (validated
like the key-level field, enterprise gated); the value is hydrated onto
UserAPIKeyAuth via the token view, TeamGrants and the carried budget state;
_check_team_model_budget enforces it in the centralized common checks; the
limiter meters spend under team_model_spend:<team>:<model>:<duration> and
skips the team counter when the key overrides; /team/update lets only a
proxy admin raise, re-window or drop a cap; /team/info exposes usage.
The Anthropic context-management compaction summary subrequest runs the
same team gate. Both fallback token-view SQL definitions project the column.
UI: team create and edit forms reuse the key-level ModelMaxBudgetEditor,
premium gated, sending {} to clear and omitting unchanged fields.
A key entry overrides the team cap only when it spend-gates the model
(non-negative max_budget); a row that only carries tpm/rpm limits or a
negative cap leaves the team cap in force.
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
The write path no longer dual-writes the global table. The cron rolls up closed UTC days
only, so a pod still flushing the current day can never leave the global table short. The
key-free arm reads days through the marker from the global table and later days from
LiteLLM_DailyUserSpend in one UNION ALL, and the marker comes from the config cache
rather than a per-request database lookup.
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Adds a daily spend table without api_key or user_id, written atomically alongside
LiteLLM_DailyUserSpend from the batched writer, reconciled from history by a
scheduled job that advances a marker in LiteLLM_Config, and read by the key-free
arm of the aggregated usage query once the marker covers the requested range.
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Auth gate rejections are raised before add_litellm_data_to_request stamps the caller User-Agent and SpendLogsMetadata dropped the field, so failure spend logs and prometheus labels could not identify an abusive client. Stamp requester_ip_address and user_agent on the failure hook payload and carry user_agent through spend log metadata. Request scopes without a headers entry are tolerated.
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Keeps the base's rule that a non-admin id lookup matching no spend-log row answers 403, so the detail route never consults cold storage without an owner row
* perf(auth): prefetch user, team, membership, org and project in one MGET, one query and one pipeline
Auth read each object with its own Redis GET and, on a miss, its own DB
query, then the admission spend counters with one GET each. The prefetch
warms every entry the checks read with one MGET, one raw query for the
Redis misses and one pipeline write, and a per-request batch serves the
spend counter reads from one MGET. The per-object getters stay the
readers and the fallback, so enforcement does not depend on the prefetch
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* refactor(auth): keep prefetch and spend batch collections immutable
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* perf(auth): let the cold spend-counter reseed reuse the admission MGET instead of one GET per counter
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* perf(auth): prefetch referenced auth objects only after the key's model access check passes
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* test(auth): give the prefetch-ordering test's patches their test-quality reasons
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* test(auth): move the real-Postgres prefetch join test to the proxy_behavior shard
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* fix(auth): read NULL nested permission and budget lists as [] in the prefetch join
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* perf(proxy): batch post-call spend counter reads and carry budget state through the request
Post-call warm checks, reservation reads and reconcile reads for one request now go through a task-local spend counter batch: one MGET answers every counter, successful increments write their result back into the batch so no second Redis read follows, and invalidation forgets the key. RedisCache.async_increment sends INCRBYFLOAT and its TTL command in one pipeline round trip.
Auth pins frozen team, user and org budget snapshots on UserAPIKeyAuth, the pre-call setup writes them into the request metadata, and Prometheus reads them back instead of calling get_key_object, get_team_object, get_user_object and get_org_object on the response path. The getters stay as the fallback for requests that carried nothing (custom auth, unauthenticated routes, skipped checks).
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* perf(proxy): reconcile the budget reservation and the post-call warm checks from one MGET and one pipeline
A scope opened inside an open spend counter batch binds into it instead of starting its own, so the reservation reconcile and the post-call warm checks share the request's single MGET. The reconcile reads every reserved counter concurrently, sends the consistent adjustments in one INCRBYFLOAT+EXPIRE pipeline and settles a flushed or reseeded counter on its own afterwards, keeping the pre-call resize fail-closed. PendingSpendIncrement moves to spend_counter_batch so budget_reservation can build a pipeline without importing a private name
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* chore(proxy): drop the dataclass import left behind by the PendingSpendIncrement move
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* fix(types): import Self from typing_extensions so the proxy imports on Python 3.10
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* test(proxy): use a neutral organization alias in the carried budget state tests
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* test(proxy): cover recorded and forgotten spend counter values in the request batch
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* test(caching): assert async_set_cache_pipeline_with_ttls keeps per-entry TTLs
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* refactor(proxy): type the reservation entry carried through reconcile adjustments
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* fix(auth): map the model table's aliases column to model_aliases in the prefetch join and read user memberships the way get_user_object does
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>
* perf(auth): prefetch user, team, membership, org and project in one MGET, one query and one pipeline
Auth read each object with its own Redis GET and, on a miss, its own DB
query, then the admission spend counters with one GET each. The prefetch
warms every entry the checks read with one MGET, one raw query for the
Redis misses and one pipeline write, and a per-request batch serves the
spend counter reads from one MGET. The per-object getters stay the
readers and the fallback, so enforcement does not depend on the prefetch
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* refactor(auth): keep prefetch and spend batch collections immutable
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* perf(auth): let the cold spend-counter reseed reuse the admission MGET instead of one GET per counter
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* perf(auth): prefetch referenced auth objects only after the key's model access check passes
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* test(auth): give the prefetch-ordering test's patches their test-quality reasons
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* test(auth): move the real-Postgres prefetch join test to the proxy_behavior shard
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* fix(auth): read NULL nested permission and budget lists as [] in the prefetch join
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* test(caching): assert async_set_cache_pipeline_with_ttls keeps per-entry TTLs
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* fix(auth): map the model table's aliases column to model_aliases in the prefetch join and read user memberships the way get_user_object does
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>
* feat(rust): count tiktoken cl100k_base admission tokens in Rust
The Rust admission token counter only had the Anthropic tokenizer, so every
other model (OpenAI gpt-4 family, Azure, Gemini, Bedrock non-Claude, Mistral)
tokenized with tiktoken on the Python inference worker.
Add an exact cl100k_base counter to litellm-token-counter: the vendored rank
file (base64 token / rank lines, the bytes Python's tiktoken uses) is parsed
into a byte-level BPE model and the cl100k split pattern is a handwritten
scanner over the shared Unicode classes, so no regex engine runs per request.
Both tokenizers share the message, tool and reply-priming accounting.
The PyO3 TokenCounter gains a from_cl100k_ranks constructor; Python reads the
rank file and passes it in, the way claude_json_str already works. The bridge
selects the counter through the same predicates litellm.token_counter uses
(huggingface_tokenizer_kind, openai_tokenizer_encoding), declines o200k_base,
downloaded HuggingFace and custom tokenizers to Python, and budget reservation
counts once per distinct tokenizer a request names.
The legacy gpt-3.5-turbo-0301 message accounting (4 per message, -1 per name)
stays in Python: the selector declines it through the predicate token_counter
itself uses.
* feat(rust): count tiktoken o200k_base admission tokens in Rust (#40794)
Add a handwritten o200k_base split scanner and TokenCounter::from_o200k_ranks
next to the cl100k_base counter, sharing MergeRanks and the request
accounting. The Python bridge selects it when openai_tokenizer_encoding
names o200k_base, so gpt-4o, gpt-4.1, gpt-5, o1/o3/o4 and chatgpt-4o
requests stop tokenizing on the Python worker under LITELLM_RUST=true
Co-authored-by: yassin <yassin@berri.ai>
---------
Co-authored-by: yassin <yassin@berri.ai>
Co-authored-by: devin-ai-integration[bot] <158243242+devin-ai-integration[bot]@users.noreply.github.com>
An out-of-range cursorless page returns nothing, and reading its total off the
offset reported more sessions than exist (page 4 of 100 sessions at page size 50
claimed 150). Only a page that holds rows, or the first page, ends the list;
anything past it falls back to the bounded count.
Claude-Session: https://claude.ai/code/session_01ESi9JwaXDww1vP3Qsrr4Mz
A cursorless page that comes back without its lookahead row is the end of the
list, so the total is offset + len(page) and the bounded grouped COUNT over the
whole spend-log table is skipped. First pages on small deployments and every
offset last page now cost one query less.
Moves the count into _count_grouped_sessions and reworks the query-optimization
test that asserted the count always runs second onto a full page, where it does.
Claude-Session: https://claude.ai/code/session_01ESi9JwaXDww1vP3Qsrr4Mz
A page size that does not divide SPEND_LOGS_PAGINATION_COUNT_CAP left the last
page starting inside the capped window and reading past it, so the rows
disagreed with the total reported next to them. The page limit now stops at the
end of that window, and has_more plus next_session_cursor still hand back a
cursor for walking further.
Claude-Session: https://claude.ai/code/session_01ESi9JwaXDww1vP3Qsrr4Mz
A page starting at or past SPEND_LOGS_PAGINATION_COUNT_CAP lies outside the
total the client is given, so it now returns no rows without running the page
query and the grouped top-N sort bound stays capped.
Rewrites the offset test to page a fake session store instead of asserting on
the generated SQL, and covers the last page inside the cap next to the first
page past it.
Claude-Session: https://claude.ai/code/session_01ESi9JwaXDww1vP3Qsrr4Mz
* fix(proxy): keep call_type and request start time on failed-request spend logs
post_call_failure_hook pops litellm_logging_obj before the failure callbacks
run, so the spend row built from request_data had a blank call_type and used
datetime.now() as the start time. A guardrail-blocked MCP tool call therefore
showed up in the Logs page as an LLM row with no call type and a 0s duration.
Lift call_type and start_time off the logging object alongside the fields
already lifted, and have the DB failure hook prefer the lifted start time.
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* test(proxy): inject the spend writer into _ProxyDBLogger instead of patching a module global
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
---------
Co-authored-by: yucheng <yucheng@berri.ai>
Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* feat(proxy): offload spend tracking to a pod-local spend worker sidecar
py-spy on the gateway showed the post-response _PROXY_track_cost_callback,
spend-log and DBSpendUpdateWriter work running on the inference workers'
event loop, so a DB or Redis stall backed up the request path.
When LITELLM_SPEND_WORKER_ENABLED=true, _ProxyDBLogger serializes one compact
typed SpendEvent per success and hands it to a SpendEventProducer that ships
it over a unix socket (default) or loopback-only TCP to a sidecar started as
`python -m gateway.spend_worker`. The sidecar runs the unchanged
_ProxyDBLogger pipeline against the pod's PgBouncer (pooled_database_url).
When the sidecar is unreachable, the buffer is full, or the gateway shuts
down with events still queued or in flight, the producer applies
LITELLM_SPEND_WORKER_ON_UNAVAILABLE (fallback in-process, or drop). The
sidecar half-closes producers on SIGTERM and drains, the producer treats
EOF as unavailable, and the gateway flushes buffered spend counters on
shutdown. The sidecar honors LITELLM_LOG so its writes are visible in its
own process log.
Helm: both charts gain an opt-in spend-worker sidecar container sharing an
emptyDir socket dir, and the componentized chart's HPA uses a
ContainerResource CPU metric scoped to the gateway container so sidecar
CPU does not drive inference scaling.
* feat(terraform): opt-in spend-worker sidecar for the AWS and GCP gateway stacks
Adds spend_worker_* inputs to both modules. On ECS Fargate the sidecar is a second, non-essential container in the gateway task; on Cloud Run it is a second container in the gateway service. Both listen on loopback TCP, share the gateway's DB/Redis/secret env, and set LITELLM_JOB_ROLE=spend_worker. Disabled by default. Plan-only tests cover both, and the terraform CI workflow now runs the gcp module too
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* test(proxy): retrieve a completed batch in the in-process spend path test
The base now defers cost tracking for batches that are still in flight, so an in_progress batch never reaches update_database
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* refactor(proxy): rename the spend worker sidecar to collector
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* fix(proxy): run the collector from the installed litellm package and finish in-flight fallbacks on shutdown
The sidecar command becomes python -m litellm.proxy.collector so the classic image, whose runtime
stage copies only the installed package, can run it. The module now assembles DATABASE_URL and the
pod-local pgbouncer URL itself, replacing gateway/collector.py
The componentized collector sidecar inherits gateway.volumeMounts so custom CA mounts reach it.
SpendEventProducer shields an in-progress fallback from the writer task cancellation so close()
no longer loses an event already handed to the in-process pipeline
Helpers used across modules (address_argument, should_store_prompts_and_responses_in_spend_logs,
flush_spend_counters_on_shutdown) become public so the change adds no reportPrivateUsage errors
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* ci(terraform): drop the gcp job duplicated by the aws/gcp matrix
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* fix(collector): keep metrics env off the classic sidecar and reject shared loopback ports
The classic chart no longer hands PROMETHEUS_METRICS_PORT and the billing metrics env to the collector container, and gives it the same /.npm scratch mount as the proxy on a read-only root. AWS and GCP now refuse a plan where the spend collector and the metrics sidecar bind the same loopback port. A regression test drives a sidecar crash mid-stream on asyncio and uvloop and checks no event is billed by both the sidecar and the in-process fallback; the producer docstring spells out why a failed drain() cannot double count
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* style(proxy): format pooled_database_url after the pgbouncer rebase
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* fix(proxy): keep the cache-hit preset key and survive dead producers on collector drain
Cache hits updated the logging object after the early return, so the offloaded spend event carried
preset_cache_key=None and the collector re-hashed reconstructed kwargs. Also guard write_eof() against
producer transports uvloop already closed so one dead connection cannot abort the drain
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* fix(terraform): keep the gcp collector port off the metrics sidecar health port
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* fix(proxy): collector connects to Postgres directly under IAM or Entra token auth
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* fix(proxy): mark the collector's DATABASE_URL as pooled when it uses the pod's pgbouncer
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>
