Conversation
| region: z.string().optional(), | ||
| startTimeUnixNano: z.number().optional(), | ||
| endTimeUnixNano: z.number().optional(), | ||
| format: z.enum(["table", "json"]).default("table"), |
There was a problem hiding this comment.
i would default to json since llm to llm prefers json only llm to user would be table
| } | ||
|
|
||
| const lines: string[] = []; | ||
| lines.push("| Trace Start Time | Root Operation | Services | Duration | Status | Trace ID |"); |
There was a problem hiding this comment.
this makes not so much sense, since spans are much more interesting to the llm and user - I would not define them as traces with trace start time but as spans - the table should hold spans, maybe grouped by trace_id - for now i would not group them but just sort them based on span start time
|
Nice work - but this is quite different in concept from what I implemented - why are you going primarily with trace as the entity to query, the llm would be more interested in the spans since there is no clear way currently to associate a trace to a action. |
2 participants
DO NOT MERGE