Commit graph

503 commits

Author SHA1 Message Date
Krish Dholakia
0cb872cbea Merge pull request #4810 from BerriAI/litellm_team_modify_guardrails
feat(auth_checks.py): Allow admin to disable team from turning on/off guardrails
2024-07-22 22:32:24 -07:00
Krrish Dholakia
14c5d6ff03 fix(proxy/utils.py): add stronger typing for litellm params in failure call logging 2024-07-22 21:31:39 -07:00
Ishaan Jaff
cd2bdfd2fb feat - return team_metadata in user_api_key_auth 2024-07-22 18:21:50 -07:00
Krrish Dholakia
a351b7cc3e feat(auth_checks.py): Allow admin to disable team from turning on/off guardrails. 2024-07-20 18:39:05 -07:00
Krrish Dholakia
b92af48854 fix(user_api_key_auth.py): update team values in token cache if refreshed more recently 2024-07-19 17:35:59 -07:00
Ishaan Jaff
75e48c84f4 fix add fix to update spend logs 2024-07-19 12:49:23 -07:00
Ishaan Jaff
3203d825e0 fix calculate correct alerting threshold 2024-07-16 21:43:17 -07:00
Ishaan Jaff
f1a3fd99b7 fix tracking hanging requests 2024-07-16 21:43:16 -07:00
Ishaan Jaff
4eef814a35 fix storing request status in mem 2024-07-16 21:43:16 -07:00
Krrish Dholakia
b022099712 fix(proxy/utils.py): fix failure logging for rejected requests. + unit tests 2024-07-16 17:15:20 -07:00
Ishaan Jaff
36b24209eb fix only use per request tracking if slack alerting is being used 2024-07-13 18:01:53 -07:00
Krrish Dholakia
1d6643df22 feat(pass_through_endpoint.py): support enforcing key rpm limits on pass through endpoints
Closes https://github.com/BerriAI/litellm/issues/4698
2024-07-13 13:29:44 -07:00
Ishaan Jaff
1adff9cbd6 Merge pull request #4684 from BerriAI/litellm_safe_memory_mode
[Feat] Allow safe memory mode
2024-07-12 18:32:16 -07:00
Ishaan Jaff
c43948545f feat add safe_memory_mode 2024-07-12 18:18:39 -07:00
Ishaan Jaff
bc7b3f28b9 reduce ttil for update_request_status 2024-07-12 15:14:54 -07:00
Ishaan Jaff
49f8894dcc fix show exact prisma exception when starting proxy 2024-07-09 18:20:09 -07:00
Krrish Dholakia
1dae0a5b6a fix(utils.py): cleanup 'additionalProperties=False' for tool calling with zod
Fixes issue with zod passing in additionalProperties=False, causing vertex ai / gemini calls to fail
2024-07-06 17:27:37 -07:00
Ishaan Jaff
f96c0efd90 Merge pull request #4576 from BerriAI/litellm_encrypt_decrypt_using_salt
[Refactor] Use helper function to encrypt/decrypt model credentials
2024-07-06 15:11:09 -07:00
Ishaan Jaff
752fe3ac7c improve sign up flow - show missing env vars 2024-07-06 13:57:19 -07:00
Krrish Dholakia
47ce6ccac0 fix(proxy_server.py): fix embedding model exception mapping 2024-07-06 11:14:41 -07:00
Ishaan Jaff
561a30dd59 move encrypt / decrypt to helper 2024-07-06 11:09:47 -07:00
Krrish Dholakia
faf11c3a3e fix(test_proxy_reject_logging.py): fix test 2024-07-05 19:09:37 -07:00
Krrish Dholakia
fe889c47db fix(utils.py): log failure to sync failure callbacks as well 2024-07-05 14:49:34 -07:00
Krrish Dholakia
deb7a86e9c fix(proxy/utils.py): support logging rejected requests to langfuse, etc. 2024-07-05 14:39:35 -07:00
Krrish Dholakia
9f039a9776 fix(proxy_server.py): fix callback check order 2024-07-05 14:06:33 -07:00
Krrish Dholakia
56410cfcd0 fix(proxy_server.py): support langfuse logging for rejected requests on /v1/chat/completions 2024-07-05 13:07:09 -07:00
Krrish Dholakia
d09a78d7fd fix(slack_alerting.py): use in-memory cache for checking request status 2024-07-02 13:01:59 -07:00
Krish Dholakia
63d0defa6d Merge branch 'main' into litellm_dynamic_tpm_limits 2024-06-22 19:14:59 -07:00
Krrish Dholakia
8843b0dc77 feat(dynamic_rate_limiter.py): working e2e 2024-06-22 14:41:22 -07:00
Krrish Dholakia
8f95381276 refactor: instrument 'dynamic_rate_limiting' callback on proxy 2024-06-22 00:32:29 -07:00
Krish Dholakia
186fc867a4 Merge pull request #4344 from BerriAI/litellm_refactor_langfuse_slack_trace_url
refactor(litellm_logging.py): refactors how slack_alerting generates langfuse trace url
2024-06-21 23:37:38 -07:00
Ishaan Jaff
aa3f2b3cf9 fix cost tracking by tags 2024-06-21 16:49:57 -07:00
Krrish Dholakia
c7b06c42b7 refactor(litellm_logging.py): refactors how slack_alerting generates langfuse trace url
gets the url from logging object
2024-06-21 16:12:25 -07:00
Krrish Dholakia
174b345766 fix(proxy/utils.py): fix add langfuse trace id to alert
Fixing the import after refactor
2024-06-21 14:55:09 -07:00
Krrish Dholakia
fb98dd70ce fix(proxy/utils.py): fix bool on check 2024-06-21 14:29:38 -07:00
Krrish Dholakia
73c2108d41 fix(proxy/utils.py): fix linting error 2024-06-20 14:13:38 -07:00
Krrish Dholakia
fa6ddcde3c fix(litellm_logging.py): fix lago callback logic 2024-06-17 09:10:19 -07:00
Krish Dholakia
fa2d8bc794 Merge pull request #4216 from BerriAI/litellm_refactor_logging
refactor(utils.py): Cut down utils.py to <10k lines.
2024-06-15 15:19:42 -07:00
Krrish Dholakia
019533d815 fix(utils.py): move 'set_callbacks' to litellm_logging.py 2024-06-15 12:02:30 -07:00
Krrish Dholakia
9d7f5d503c refactor(utils.py): refactor Logging to it's own class. Cut down utils.py to <10k lines.
Easier debugging

 Reference: https://github.com/BerriAI/litellm/issues/4206
2024-06-15 10:57:20 -07:00
Ishaan Jaff
4bc2bfb176 fix - proxy refactor user_api_key_auth 2024-06-15 10:33:58 -07:00
Ishaan Jaff
8988b2e909 Merge pull request #4209 from BerriAI/litellm_send_email_alerts_budget_exceeded
[Feat] send email alerts when budget exceeded
2024-06-14 20:23:19 -07:00
Ishaan Jaff
614f41d12e fix -better debugging before sending emails 2024-06-14 17:38:33 -07:00
Krrish Dholakia
1cce99300f fix(slack_alerting.py): allow new 'alerting_metadata' arg
Allows user to pass in additional alerting metadata for debugging
2024-06-14 16:06:47 -07:00
Krrish Dholakia
af2aeb595d fix(proxy/utils.py): fix reset monthly budget
fix to reset at the same time each month (not at start of month)
2024-06-14 14:41:06 -07:00
Krish Dholakia
c373f104cc Merge branch 'main' into litellm_redis_cache_usage 2024-06-13 22:07:21 -07:00
Krrish Dholakia
417e25ae08 feat(proxy/utils.py): allow budget duration in months
Closes https://github.com/BerriAI/litellm/issues/4042
2024-06-13 16:52:17 -07:00
Krrish Dholakia
77328e4a28 fix(parallel_request_limiter.py): use redis cache, if available for rate limiting across instances
Fixes https://github.com/BerriAI/litellm/issues/4148
2024-06-12 10:35:48 -07:00
Krrish Dholakia
43e85a3993 fix(proxy/_types.py): support logging k,v pairs to spend logs with spend_logs_metadata param 2024-06-12 08:42:35 -07:00
Ishaan Jaff
1b73da1b21 fix refactor management endpoint utils 2024-06-11 16:16:10 -07:00