docs(dynamo-run): Remove vllm/sglang/trtllm engines from dynamo-run docs (#2332)

launch/dynamo-run/src/flags.rs

@@ -74,7 +74,6 @@ pub struct Flags {
 
     /// Maximum number of batched tokens for KV routing
     /// Needed for informing the KV router
-    /// TODO: derive from vllm args
     /// NOTE: this is not actually used for now
     #[arg(long, default_value = "8192")]
     pub max_num_batched_tokens: Option<u32>,
@@ -103,10 +102,11 @@ pub struct Flags {
     #[arg(long)]
     pub context_length: Option<u32>,
 
-    /// KV cache block size (vllm only)
+    /// KV cache block size (is this used? Maybe by Python vllm worker?)
     #[arg(long)]
     pub kv_cache_block_size: Option<u32>,
 
+    /// Mocker engine only.
     /// Additional engine-specific arguments from a JSON file.
     /// Contains a mapping of parameter names to values.
     #[arg(long)]
@@ -203,6 +203,16 @@ impl Flags {
                 // nothing to check here
             }
         }
+
+        match out_opt {
+            Output::Mocker => {}
+            _ => {
+                if self.extra_engine_args.is_some() {
+                    anyhow::bail!("`--extra-engine-args` is only for the mocker engine");
+                }
+            }
+        }
+
         Ok(())
     }
 

launch/dynamo-run/src/main.rs

@@ -3,7 +3,7 @@
 
 use std::env;
 
-use clap::Parser;
+use clap::{CommandFactory as _, Parser};
 
 use dynamo_llm::entrypoint::input::Input;
 use dynamo_run::Output;
@@ -22,9 +22,11 @@ Example:
 - cd target/debug
 - ./dynamo-run Qwen/Qwen3-0.6B
 - OR: ./dynamo-run /data/models/Llama-3.2-1B-Instruct-Q4_K_M.gguf
+
+See `docs/guides/dynamo_run.md` in the repo for full details.
 "#;
 
-const USAGE: &str = "USAGE: dynamo-run in=[http|text|dyn://<path>|batch:<folder>] out=ENGINE_LIST|auto|dyn://<path> [--http-port 8080] [--model-path <path>] [--model-name <served-model-name>] [--model-config <hf-repo>] [--tensor-parallel-size=1] [--context-length=N] [--kv-cache-block-size=16] [--num-nodes=1] [--node-rank=0] [--leader-addr=127.0.0.1:9876] [--base-gpu-id=0] [--extra-engine-args=args.json] [--static-worker] [--router-mode random|round-robin|kv] [--kv-overlap-score-weight=2.0] [--kv-gpu-cache-usage-weight=1.0] [--kv-waiting-requests-weight=1.0] [--migration-limit=0] [--verbosity (-v|-vv)]";
+const USAGE: &str = "USAGE: dynamo-run in=[http|text|dyn://<path>|batch:<folder>] out=ENGINE_LIST|auto|dyn://<path> [--http-port 8080] [--model-path <path>] [--model-name <served-model-name>] [--model-config <hf-repo>] [--context-length=N] [--kv-cache-block-size=16] [--extra-engine-args=args.json] [--static-worker] [--router-mode random|round-robin|kv] [--kv-overlap-score-weight=2.0] [--router-temperature=0.0] [--use-kv-events] [--max-num-batched-tokens=1.0] [--migration-limit=0] [--verbosity (-v|-vv)]";
 
 fn main() -> anyhow::Result<()> {
     // Set log level based on verbosity flag
@@ -71,6 +73,7 @@ async fn wrapper(runtime: dynamo_runtime::Runtime) -> anyhow::Result<()> {
         let usage = USAGE.replace("ENGINE_LIST", &engine_list);
         println!("{usage}");
         println!("{HELP}");
+        dynamo_run::Flags::command().print_long_help().unwrap();
         return Ok(());
     } else if args[0] == "--version" {
         if let Some(describe) = option_env!("VERGEN_GIT_DESCRIBE") {