[Bug]: image tag main-stable got reverted to v1.77 (previously v1.79)

[lephuongbg]

What happened?

Not sure if this was intentional, but I was rebuilding my own image from main-stable and discovered that litellm version got reverted to v1.77, which causedd issue when deployment get rolled out due to max_requests_before_restart argument does not exist on older version.

Relevant log output

Are you a ML Ops Team?

No

What LiteLLM version are you on ?

v1.79.3

Twitter / LinkedIn details

No response

[21y4d]

I also faced this issue. My models stopped loading because litellm showed 0 models available. Upon checking litellm logs, it was returning encryption related errors:

iteLLM Proxy:ERROR: encrypt_decrypt_utils.py:63 - Error decrypting value for key: litellm_credential_name, Did your master_key/salt key change recently? 

I temporarily changed my docker image to get it working again:

#image: ghcr.io/berriai/litellm:main-stable
image: ghcr.io/berriai/litellm:v1.79.3-stable

The main-stable tag reverted to an older version, which appears to have caused issues as litellm configs/db didn't like being reverted to older versions.

[mertcangokgoz]

Yes, while the version in main-stable image was 1.79, it's now 1.77.