scriptoriumadmin/litellm
1
0
Fork 0
mirror of https://github.com/BerriAI/litellm.git synced 2026-10-09 03:18:44 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 108

fix(purview): use post-await timestamp for cache TTL

Browse source 
Capture the timestamp after the network call completes when storing it
as the cache freshness marker, so the effective TTL reflects when the
response was actually received rather than when the request started.
Under high network latency the previous behavior shortened the
effective cache lifetime.

Co-authored-by: Yassin Kortam <yassin@berri.ai>
This commit is contained in:
Cursor Agent 2026-05-22 16:46:29 +00:00
parent 30b5e820f1
commit 5876b826a8
No known key found for this signature in database
1 changed files with 7 additions and 2 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

9
litellm/proxy/guardrails/guardrail_hooks/microsoft_purview/base.py
View file

@ -110,8 +110,10 @@ class PurviewGuardrailBase:
token_data = response.json()
access_token = token_data["access_token"]
expires_in = int(token_data.get("expires_in", 3599))
# Recompute ``now`` after the await so the expiry reflects when the
# token was actually received, not when the request started.
with self._cache_lock:
self._token_cache = (access_token, now + expires_in)
self._token_cache = (access_token, time.time() + expires_in)
verbose_proxy_logger.debug(
"Purview: acquired new OAuth2 token (expires_in=%ds)", expires_in
)
@ -188,8 +190,11 @@ class PurviewGuardrailBase:
response_json, response_headers = await self._graph_post(url, body)
etag = response_headers.get("etag", response_headers.get("ETag", ""))
# Recompute ``now`` after the await so the TTL reflects when the
# scope response was actually received, not when the request started.
fetched_at = time.time()
with self._cache_lock:
self._scope_cache[user_id] = (etag, response_json, now)
self._scope_cache[user_id] = (etag, response_json, fetched_at)
# Move refreshed entry to the end so it is treated as most-recently-used.
# OrderedDict.__setitem__ preserves existing insertion order for known
# keys, so an explicit move_to_end() call is required.