I am trying to access Llama 3.1 API service on vertex AI via Portkey. followed this documentation - https://docs.portkey.ai/docs/welcome/integration-guides/vertex-ai to make a virtual key but i don't see llama on the portkey playground. have enabled the llama api service too. could someone help me debug this? cc

I got an error from OpenAI:


if i look at the portkey dashboard, i see 824K tokens for the entire day. can you help me debug this?

the portkey api 1.1.x -> 1.2.x had a bunch of breaking changes, do you guys share breaking changes somewhere?

@Vrushank | Portkey - the replay button seems to have gone yet again 😐

The replay button seems to have gone again

The tool call support on replay isn't really working. i'm getting an error openai error: Missing parameter 'tool_call_id': messages with role 'tool' must have a 'tool_call_id'. and no option to add a tool_call_id on the tool message. when can we expect to have tool call support working for open ai? the limtation is really slowing us down.

Round Table Legal AI Inc.

i noticed a bug when using the "reply" feature on a log which had tool calls and tool messages in the history. the tool call is not being parsed as expected, only a tool message shows up and i get an error saying "tool message should follow a tool call". the tool call showed up as a blank assistant message.

I read on the documentation that on virtual keys, we can assign limits to cost. i want to set virtual keys for potential candidates i am looking to hire and set a value limit of $1. how can i do that?

when i am working with the playground, the "an update is available" button keeps popping up. even after i hit update, it'll come again in a few seconds. and if i click on it, all progress on the playground is lost

one more thing, while the above might take time, is it possible to provide a copy button beside each message in logs? that'll make it easier to transfer a conversation into a prompt playground

@Team Portkey - i'm running into a wierd issue where i am instructing open ai to return a json response. i'd expect a 100-200 word response. but when open ai starts making the json and ends abruptly after 2 lines, giving an incomplete and incorrectly formatted json. strange thing is that i have a retry set and it gives the exact same response 5 times in a row. i dont have any caching enabled, could this be an issue on portkey side?

@Team Portkey - i tried a config with timeout of 30s and 3 retries. i doesn't look it does an automatic retry if the timeout happens. the retry is only triggered on failure. is this expected behaviour?

@Team Portkey - i tried the anthropic models using portkey today. i keep getting an internal server error - 05 when i use anthropic, but it works on open ai models. is it possible for you guys to propogate the error message back to the user interface?