From 8b5dcc8d813c0c08202f232ff68c4eaef4e2d14f Mon Sep 17 00:00:00 2001
From: narmstrong-berry <narmstrong@berri.ai>
Date: Wed, 21 Oct 2026 06:09:32 +0000
Subject: [PATCH] feat(loadtest): sustained-250 workload

---
 litellm-rust/loadtest/README.md               |  2 +-
 .../loadtest/workloads/sustained-250.json     | 20 +++++++++++++++++++
 2 files changed, 21 insertions(+), 1 deletion(-)
 create mode 100644 litellm-rust/loadtest/workloads/sustained-250.json

diff --git a/litellm-rust/loadtest/README.md b/litellm-rust/loadtest/README.md
index 52e3e10157b..16c5fd3b73c 100644
--- a/litellm-rust/loadtest/README.md
+++ b/litellm-rust/loadtest/README.md
@@ -16,7 +16,7 @@ machines.
 | `s3stub` | In-memory path-style S3 (PUT, GET, ListObjectsV2) for the `s3_v2` logs of a full-feature run. Objects are stored compressed and listed from a sorted index, so millions of logs fit. |
 
 Workloads live in `workloads/`: `smoke`, `realistic-ramp`, `long-streams`, `large-context`,
-`key-churn` and `soak`. `sustained-100`, `sustained-500` and `sustained-1000` hold one rate for
+`key-churn` and `soak`. `sustained-100`, `sustained-250`, `sustained-500` and `sustained-1000` hold one rate for
 15 minutes with evenly spaced arrivals and smoke's request mix over every model group;
 `sustained-check` runs the same mix at 100, 500 and 1,000 rps for 2 minutes each, to check that the
 client, mock and stack hold 1,000 rps before the long runs.
diff --git a/litellm-rust/loadtest/workloads/sustained-250.json b/litellm-rust/loadtest/workloads/sustained-250.json
new file mode 100644
index 00000000000..64d43648e80
--- /dev/null
+++ b/litellm-rust/loadtest/workloads/sustained-250.json
@@ -0,0 +1,20 @@
+{
+  "name": "sustained 250 rps for 15 minutes, evenly spaced arrivals, smoke's request mix over every model group",
+  "stages": [{"duration_s": 900, "rps": 250}],
+  "arrivals": "constant",
+  "nonstream_max_s": 360,
+  "warmup_s": 60,
+  "max_in_flight": 30000,
+  "keys": {"file": "out/keys.txt", "zipf_s": 0},
+  "models": {"chat": {"prefix": "gpt-6-astra-", "count": 10}, "messages": {"prefix": "claude-opus-5-5-", "count": 10}},
+  "classes": [
+    {"name": "chat", "weight": 1, "endpoint": "chat", "stream_fraction": 0.5,
+     "context_tokens": {"dist": "loguniform", "min": 10000, "max": 50000},
+     "ttft_ms": {"dist": "uniform", "min": 200, "max": 800},
+     "duration_s": {"dist": "uniform", "min": 1, "max": 5}},
+    {"name": "messages", "weight": 1, "endpoint": "messages", "stream_fraction": 0.5,
+     "context_tokens": {"dist": "loguniform", "min": 10000, "max": 50000},
+     "ttft_ms": {"dist": "uniform", "min": 200, "max": 800},
+     "duration_s": {"dist": "uniform", "min": 1, "max": 5}}
+  ]
+}