CLIENT PAUSE ALL for the length of the chaos phase instead of CLIENT PAUSE WRITE, so every Redis touchpoint on the request path times out rather than just the writes. The pause is sized to the phase because it freezes the control connection too; teardown's CLIENT UNPAUSE is a safety net for a phase that overran
Latency, RSS and CPU are now budgeted as chaos-over-baseline ratios (p50/p90/p99 for latency and RSS, CPU seconds per request once) through a small phase_budget module, replacing the machine-shaped absolutes. The Redis timeout rate is reported but no longer asserted
The final /metrics scrape waits for litellm_deployment_failure_responses_total to stop moving, since that counter is bumped from the async logging queue and lagged the load generator by thousands of increments. The model group carries a unique marker so a deployment left behind by an aborted run cannot absorb this run's retries
Co-Authored-By: Claude Code <noreply@anthropic.com>
The docstring cited session_affinity_ttl_seconds as the keepalive bound,
but the Router-level knob feeding ttl_seconds is
deployment_affinity_ttl_seconds; session_affinity_ttl_seconds is the
separate per-request PreRoutingHookResponse override. Anyone grepping
the docstring's name to shrink the Router default finds only the
override. Name both, scoped correctly.
Starlette scans the route table in registration order, so a request pays one
regex match per route registered ahead of its own. The proxy registers several
hundred routes and left the liveness probe near position 280 and the lazy
loaded /v1/messages at the very end. Move /health/liveliness, /health/liveness,
/v1/chat/completions, /chat/completions and /v1/messages to the front of the
route table after startup registration and again after a lazy router loads.
Co-authored-by: yassin <yassin@berri.ai>
Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* feat(proxy): make the in-memory management cache capacity configurable
Add general_settings.user_api_key_cache_max_size (positive int, default 200) to resize the
in-memory tier of the shared user_api_key_cache at startup and on DB config reloads, expose it
in the Admin UI general settings, and cover it with behavioral tests. Prior art: #34726
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* fix(caching): resize the in-memory tier from DualCache so any cache instance honours the cap
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* style(proxy): wrap the cache capacity field description to the 120 col limit
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>
* fix(passthrough): parse Bedrock stream spend incrementally instead of buffering the whole response
Bedrock pass-through streaming kept every relayed chunk in memory until EOF and
then decoded, parsed and translated the whole stream again for spend logging.
Large or concurrent streams could exhaust proxy worker memory.
Sync and async passthrough wrappers now hand each chunk to a provider stream
collector as it is relayed. Bedrock decodes event-stream frames incrementally,
folds consecutive text deltas, and keeps only what stream_chunk_builder needs
for usage, tool calls and metadata. Text deltas are no longer retained in the
Bedrock and Anthropic stream decoders either. Providers without a collector
keep the previous raw-bytes behavior. Collector failures are isolated so spend
tracking can never interrupt the customer stream
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* test(passthrough): assert the spend payload the collector builds instead of mock internals
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* test(passthrough): type the Bedrock collector helpers by the collector protocol instead of asserting the class
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>
Key objects share the 200-entry UserApiKeyCache in-memory store with teams,
end users, tags and memberships, so churn in those objects evicts hot keys
and forces a LiteLLM_VerificationToken lookup on the next request. Route
bare hashed-token keys to a dedicated InMemoryCache inside UserApiKeyCache
while keeping Redis, TTL, serialization and invalidation shared
Co-authored-by: yassin <yassin@berri.ai>
Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
The sequential version sent one request at a time, so a Redis outage never
reached the concurrency where the failed-tracking alert body actually grows.
This drives the proxy with locust against one model group of three mock
deployments, two failing at order 1 and one serving at order 2, so every
request spends its retries on the failing pair and lands on the serving
deployment through the order-based fallback. Two phases, a healthy baseline
and a CLIENT PAUSE WRITE window, and every request must succeed in both.
Latency, RSS and CPU are reported as p50/p90/p99 per phase rather than
asserted on: RSS and CPU come from psutil on the proxy's process tree, since
a multi-worker proxy serves /metrics from the prometheus multiprocess
collector and that drops the process collector's series. Thresholds stay open
until weekly runs give real baselines.
Co-Authored-By: Claude Code <noreply@anthropic.com>
The provider budget push runs inside the request success callback, so
awaiting the Redis pipeline there made every request wait for the round
trip. Hand it back to a task whose failure is logged through the breaker
aware logger, so an open breaker stays a debug line and a real Redis error
is one error line instead of an unretrieved task traceback
The componentized chart rendered litellm.connectionPoolEnv into the gateway
container only, so with database.connectionPool.enabled and
gateway.collector.enabled the collector's Prisma client opened its own pool
straight to Postgres instead of going through the pod-local PgBouncer.
Render the same include in the collector container, drop the gateway.extraEnv
workaround from the collector test, and add enabled/disabled regression
assertions for the collector in both helm charts and the terraform aws and gcp
modules, which already pass the pool env to their collector containers. The
module READMEs note the IAM token-auth exception, where the collector keeps a
direct connection by design.
Co-authored-by: yassin <yassin@berri.ai>
Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
The shared user popover copied alias, email and ID through three copy
buttons that all announced themselves as "Copy ID", so a screen reader
could not tell them apart. IdCell now takes the label, defaulting to the
old text everywhere else.
Also drops the closest("tr") the new link tests used, which put the
testing-library/no-node-access budget over its ceiling, and asserts the
sentinel row leaves User Email and the admin badge unlinked too.
Claude-Session: https://claude.ai/code/session_01NfwfQhamRNnSqgXMUjf3h4
The team detail page's Virtual Keys table showed Organization ID, User
Email, User ID and Created By as dead text, so getting from a key to the
org or user behind it meant copying an id and searching for it.
Those four cells now render as links, reusing the sentinel-aware href
helpers, so default_user_id and the litellm-dashboard team stay plain
text instead of pointing at pages that do not exist.
The Created By cell was a verbatim copy of the Virtual Keys page's user
popover, so that moved into the shared table_cells kit and both tables
now use the one implementation.
Claude-Session: https://claude.ai/code/session_01NfwfQhamRNnSqgXMUjf3h4
* fix(streaming): keep usage-only chunks from crashing streams with empty stream_options
The usage-only chunk branch in CustomStreamWrapper.chunk_creator indexed stream_options["include_usage"] directly, so a caller passing stream_options={} hit a KeyError that surfaced as MidStreamFallbackError. Streaming mock_response with an admission input_tokens count (#40637) now always emits such a chunk, which made the crash reachable. Reuse the send_stream_usage policy computed at init instead. Also annotate the #40637 test bindings with Final.
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* fix(streaming): report admitted zero prompt tokens instead of recounting in mock streams
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>
