[CI] Add retry with 4x backoff to HTTP fetches for transient failures (#37218)

Signed-off-by: Andreas Karatzas <akaratza@amd.com>
This commit is contained in:
Andreas Karatzas
2026-03-19 19:00:20 +00:00
committed by GitHub
parent e5d96dc8fc
commit fb8b5e05fc
2 changed files with 225 additions and 13 deletions
+7
View File
@@ -64,6 +64,7 @@ if TYPE_CHECKING:
VLLM_IMAGE_FETCH_TIMEOUT: int = 5
VLLM_VIDEO_FETCH_TIMEOUT: int = 30
VLLM_AUDIO_FETCH_TIMEOUT: int = 10
VLLM_MEDIA_FETCH_MAX_RETRIES: int = 3
VLLM_MEDIA_URL_ALLOW_REDIRECTS: bool = True
VLLM_MEDIA_LOADING_THREAD_COUNT: int = 8
VLLM_MAX_AUDIO_CLIP_FILESIZE_MB: int = 25
@@ -773,6 +774,11 @@ environment_variables: dict[str, Callable[[], Any]] = {
"VLLM_AUDIO_FETCH_TIMEOUT": lambda: int(
os.getenv("VLLM_AUDIO_FETCH_TIMEOUT", "10")
),
# Maximum number of retries for fetching media (images, audio, video)
# from URLs. Each retry quadruples the timeout. Default is 3.
"VLLM_MEDIA_FETCH_MAX_RETRIES": lambda: int(
os.getenv("VLLM_MEDIA_FETCH_MAX_RETRIES", "3")
),
# Whether to allow HTTP redirects when fetching from media URLs.
# Default to True
"VLLM_MEDIA_URL_ALLOW_REDIRECTS": lambda: bool(
@@ -1768,6 +1774,7 @@ def compile_factors() -> dict[str, object]:
"VLLM_IMAGE_FETCH_TIMEOUT",
"VLLM_VIDEO_FETCH_TIMEOUT",
"VLLM_AUDIO_FETCH_TIMEOUT",
"VLLM_MEDIA_FETCH_MAX_RETRIES",
"VLLM_MEDIA_URL_ALLOW_REDIRECTS",
"VLLM_MEDIA_LOADING_THREAD_COUNT",
"VLLM_MAX_AUDIO_CLIP_FILESIZE_MB",