Pressing c on the event picker gives this information about the BYTES column:
bytes an opaque tunnel carried: ↑ sent, ↓ received
I asked to see that column because I received a message from a version of Claude I configured to use a non-standard model:
⏺ API Error: 400 litellm.ContextWindowExceededError: litellm.BadRequestError: ContextWindowExceededError: Hosted_vllmException - {"error":{"message":"This model's maximum context length is 275000 tokens. However, you requested 1 output tokens and your prompt contains at least 275000 input tokens, for a total of at least 275001 tokens. Please reduce the length of the input prompt or the number of requested output tokens. (parameter=input_tokens, value=275000)","type":"BadRequestError","param":"input_tokens","code":400}}
model=rits/zai-org/glm-5-3. context_window_fallbacks=None. fallbacks=None.
I was curious how large a request was sent to LiteLLM. (The request was 121406 bytes). I enabled the BYTES column of the event picker but there was no content.
To see the request size that triggered the error I had to y Yank it and then switch to another window and check the size of the yanked file. Perhaps every event should have optional BYTES count, not just opaque tunnel sizes.
Pressing
con the event picker gives this information about the BYTES column:bytes an opaque tunnel carried: ↑ sent, ↓ receivedI asked to see that column because I received a message from a version of Claude I configured to use a non-standard model:
I was curious how large a request was sent to LiteLLM. (The request was 121406 bytes). I enabled the BYTES column of the event picker but there was no content.
To see the request size that triggered the error I had to
yYank it and then switch to another window and check the size of the yanked file. Perhaps every event should have optional BYTES count, not just opaque tunnel sizes.