Swapping models can take a long time and leave a lot of silence while the model is loading. Rather than silently load the model in the background, this PR allows llama-swap to send status updates in the reasoning_content of a streaming chat response.
Fixes: #366
* proxy: refactor metrics recording
- remove metrics_middleware.go as this wrapper is no longer needed. This
also eliminiates double body parsing for the modelID
- move metrics parsing to be part of MetricsMonitor
- refactor how metrics are recording in ProxyManager
- add MetricsMonitor tests
- improve mem efficiency of processStreamingResponse
- add benchmarks for MetricsMonitor.addMetrics
- proxy: refactor MetricsMonitor to be more safe handling errors
add a wake-on-lan proxy for llama-swap. When the target llama-swap server is unreachable it will send hold a request, send a WoL packet and proxy the request when llama-swap is available.
Changes:
- add Metadata key to ModelConfig
- include metadata in /v1/models under meta.llamaswap key
- add recursive macro substitution into Metadata
- change macros at global and model level to be any scalar type
Note:
This is the first mostly AI generated change to llama-swap. See #333 for notes about the workflow and approach to AI going forward.
* proxy/config: create config package and migrate configuration
The configuration is become more complex as llama-swap adds more
advanced features. This commit moves config to its own package so it can
be developed independently of the proxy package.
Additionally, enforcing a public API for a configuration will allow
downstream usage to be more decoupled.
This adds a new API endpoint, /api/models/unload/*model, that unloads a single model. In the UI when a model is in a ReadyState it will have a new button to unload it.
Fixes#312
The upstream handler would break on model IDs that contained a forward
slash. Model IDs like "aaa/bbb" called at upstream/aaa/bbb would result
in an error. This commit adds support for model IDs with a forward slash
by iteratively searching the path for a match.
Fixes: #229
Add barebones but working implementation of model preload
* add config test for Preload hook
* improve TestProxyManager_StartupHooks
* docs for new hook configuration
* add a .dev to .gitignore
* Decouple MetricsMiddleware from downstream handlers
Remove ls-real-model-name optimization. Within proxyOAIHandler the
request body's bytes are required for various rewriting features
anyways. This negated any benefits from trying not to parse it twice.
* Add endpoint aliases for reranking models
* Add MetricsMiddleware to the previous reranking endpoint
* Fix the embeddings endpoint not having model set
Major internal refactor to use an event bus to pass event/messages along. These changes are largely invisible user facing but sets up internal design for real time stats and information.
- `--watch-config` logic refactored for events
- remove multiple SSE api endpoints, replaced with /api/events
- keep all functionality essentially the same
- UI/backend sync is in near real time now
llama-swap can strip specific keys in JSON requests. This is useful for removing the ability for clients to set sampling parameters like temperature, top_k, top_p, etc.
Refactor internal upstream process life cycle management to recover better from unexpected situations. With this change llama-swap should never need to be restarted due to a crashed upstream child process. The `StateFailed` state was removed in favour of always trying to start/restart a process.
Sometimes upstreams can accept HTTP but never respond causing requests
to build up waiting for a response. This can block Process.Stop() as
that waits for inflight requests to finish. This change refactors the
code to not wait when attempting to shutdown the process.
Groups allows more control over swapping behaviour when a model is requested. The new groups feature provides three ways to control swapping: within the group, swapping out other groups or keep the models in the group loaded persistently (never swapped out).
Closes#96, #99 and #106.
This change revamps the internal logging architecture to be more flexible and descriptive. Previously all logs from both llama-swap and upstream services were mixed together. This makes it harder to troubleshoot and identify problems. This PR adds these new endpoints:
- `/logs/stream/proxy` - just llama-swap's logs
- `/logs/stream/upstream` - stdout output from the upstream server
Changes to CORS functionality:
- `Access-Control-Allow-Origin: *` is set for all requests
- for pre-flight OPTIONS requests
- specify methods: `Access-Control-Allow-Methods: GET, POST, PUT, PATCH, DELETE, OPTIONS`
- if the client sent `Access-Control-Request-Headers` then echo back the same value in `Access-Control-Allow-Headers`. If no `Access-Control-Request-Headers` were sent, then send back a default set
- set `Access-Control-Max-Age: 86400` to that may improve performance
- Add CORS tests to the proxy-manager
* Adds an endpoint '/running' that returns either an empty JSON object if no model has been loaded so far, or the last model loaded (model key) and it's current state (state key). Possible state values are: stopped, starting, ready and stopping.
* Improves the `/running` endpoint by allowing multiple entries under the `running` key within the JSON response.
Refactors the `/running` method name (listRunningProcessesHandler).
Removes the unlisted filter implementation.
* Adds tests for:
- no model loaded
- one model loaded
- multiple models loaded
* Adds simple comments.
* Simplified code structure as per 250313 comments on PR #65.
---------
Co-authored-by: FGDumitru|B <xelotx@gmail.com>
The profile slug in a model name, `profile:model`, is specific to
llama-swap. This strips `profile:` out of the model name request so
upstreams that expect just `model` work and do not require knowing about
the profile slug.
Introduce `Process.Shutdown()` and `ProxyManager.Shutdown()`. These two function required a lot of internal process state management refactoring. A key benefit is that `Process.start()` is now interruptable. When `Shutdown()` is called it will break the long health check loop.
State management within Process is also improved. Added `starting`, `stopping` and `shutdown` states. Additionally, introduced a simple finite state machine to manage transitions.
A panic occurs when a request for an invalid profile:model pair is made.
The edge case is that the profile exists and the model exists but they're
not configured as a pair.
This adds an additional check to make sure the profile:model pair is
valid before attempting to swap the model.
