[Core][Metrics] Remove vllm:prompt_tokens_recomputed metric (#38709)

Signed-off-by: Mark McLoughlin <markmc@redhat.com>
This commit is contained in:
Mark McLoughlin
2026-04-12 12:22:01 +03:00
committed by GitHub
parent ee3c0c83db
commit 72ff142c37
3 changed files with 6 additions and 27 deletions
+3 -5
View File
@@ -190,8 +190,7 @@ def test_prompt_token_stats_full_local_cache_recompute():
)
assert stats.computed == 1
assert stats.local_cache_hit == 1000
assert stats.recomputed_tokens == 1
assert stats.local_cache_hit == 999
def test_prompt_token_stats_full_external_transfer_recompute():
@@ -201,11 +200,10 @@ def test_prompt_token_stats_full_external_transfer_recompute():
# Case 6: Full external transfer (999 cached after reduction, 1 recomputed)
stats.update_from_output(
num_cached_tokens=999,
num_external_computed_tokens=1000,
num_external_computed_tokens=999,
prompt_len=1000,
)
assert stats.computed == 1
assert stats.local_cache_hit == 0
assert stats.external_kv_transfer == 1000
assert stats.recomputed_tokens == 1
assert stats.external_kv_transfer == 999