tiennm99/litellm
1
0
Fork 0
mirror of https://github.com/tiennm99/litellm.git synced 2026-08-10 16:22:17 +00:00
Code Issues Packages Projects Releases Wiki Activity

docs: fix DEFAULT_NUM_WORKERS_LITELLM_PROXY default (1, not 4) (#21127)

Browse Source 
- config_settings.md: state default is 1, clarify NUM_WORKERS recommendation
- proxy_cli.py: correct --num_workers help text to match actual default
This commit is contained in:
Alexsander Hamir
2026-02-13 07:36:20 -08:00
committed by GitHub
parent e17c639fb1
commit 68122ede2c
2 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/my-website/docs/proxy/config_settings.md
+1 -1
View File
@@ -555,7 +555,7 @@ router_settings:
| DEFAULT_MOCK_RESPONSE_COMPLETION_TOKEN_COUNT | Default token count for mock response completions. Default is 20
| DEFAULT_MOCK_RESPONSE_PROMPT_TOKEN_COUNT | Default token count for mock response prompts. Default is 10
| DEFAULT_MODEL_CREATED_AT_TIME | Default creation timestamp for models. Default is 1677610602
| DEFAULT_NUM_WORKERS_LITELLM_PROXY | Default number of workers for LiteLLM proxy. Default is 4. **We strongly recommend setting NUM Workers to Number of vCPUs available**
| DEFAULT_NUM_WORKERS_LITELLM_PROXY | Default number of workers for LiteLLM proxy when `NUM_WORKERS` is not set. Default is 1. **We strongly recommend setting NUM_WORKERS to the number of vCPUs available** (e.g. `NUM_WORKERS=8` or `--num_workers 8`).
| DEFAULT_PROMPT_INJECTION_SIMILARITY_THRESHOLD | Default threshold for prompt injection similarity. Default is 0.7
| DEFAULT_POLLING_INTERVAL | Default polling interval for schedulers in seconds. Default is 0.03
| DEFAULT_REASONING_EFFORT_DISABLE_THINKING_BUDGET | Default reasoning effort disable thinking budget. Default is 0
litellm/proxy/proxy_cli.py
+1 -1
View File
@@ -318,7 +318,7 @@ class ProxyInitializationHelpers:
@click.option(
"--num_workers",
default=DEFAULT_NUM_WORKERS_LITELLM_PROXY,
help="Number of uvicorn / gunicorn workers to spin up. By default, it equals the number of logical CPUs in the system, or 4 workers if that cannot be determined.",
help="Number of uvicorn / gunicorn workers to spin up. Default is 1 (from DEFAULT_NUM_WORKERS_LITELLM_PROXY)",
envvar="NUM_WORKERS",
)
@click.option("--api_base", default=None, help="API base URL.")