Description
Metrics discussion
Summary: lets think about what metrics we want to capture from the upstream perspective
Background
Relates to: #281
Our metrics infra will be entirely up soon. The deployments went through in #391 with some minor follow on fixes. From the UI side, we can leverage the default metrics in Umami with the implementation of their tracking codes in our TS codebase (#411). As a result, I would like to start a discussion on what custom pieces we hope to capture.
What Umami gives us out of the box
To see what Umami provides in depth please refer to their docs on the subject.
What metrics we would like to gather so far
Note: this will exclude some features that might take a while to implement!
Features excluded:
- auto-QNA generation and context chunking
- knowledge document propulating fields accuracy
Chat:
- how many requests are hitting the chatbot (per user, agregate, etc.)
Submissions (Primarily knowledge but skills where applicable):
- count / percentage of complemtions that get through each contribution stage (Author Information, Knowledge information)
- time per submission
- Author information:
- accuracy prepopulating fields (If Applicable)
- knowledge information:
- number of contributers per submission
- File Path Information:
- Some way to track if users are using good document paths?
- Document Information:
- How many manually enter documnets vs automatically upload documents
- knowledge documents upload count, count per use
- number of issues with md knowledge document upload, error count and type
- Attribution details:
- accuracy prepopulating fields
- how often users accesses the help and what fields they are having issues with
- what licenseses the are being used
Stakeholders to engage
@vishnoianil
@nerdalert
@andybraren
@Misjohns
@aevo98765
@RobotSail
@toraponibm
And anyone else I may have missed.
Metadata
Metadata
Assignees
Labels
Type
Projects
Status