POST
/
api
/
external
/
list-warnings

You will get the full list of warnings generated by the platform. Warnings are splitted in two main categories:

  • Business issues: warnings affecting the business, like a customer complaining about the pricing.
  • Model warnings, i.e. the issues related to poor LLM performance.

You need an API key to authenticate. See here for more information about the API keys.

time_range
object
required

The time range to consider when computing the metrics and analytics to show.

kind
string
required

The value of the time range. It can be one between today, 7_days, this_month, 30_days, 6_months, 12_months, ytd, custom. If custom is given the values pass in the start and end fields will be used as start date and end date if the time range.

start
string
required

The start date for the time-range. Please consider that this parameter is used only when custom is given as kind.

end
string
required

The end date for the time-range. Please consider that this parameter is used only when custom is given as kind.

filters
array

Filters you can apply to the warnings. Available filters are:

  • user: the end user of your product
  • language: the language of the model response
  • tags: custom tags you defined from the SDK Here is an example of filters called with all the value above:
{
  "filters":[
    {
      "column":"tag:custom_tag",
      "values":["custom_value"]
    },
    {
      "column":"language",
      "values":["English"]
    },
    {
      "column":"user",
      "values":["test"]
    }
  ]
}
type
string

You can use this parameter to select either the business or model warnings. If not given both warnings types are returned.

limit
integer

Parameter needed for pagination. It indicates the number of intents per page.

offset
integer

Parameter needed for pagination. It indicates the offset to be applied to the first intents returned.