Provide a metric exposing number of packages that were requested by a user but were not found as solved

[fridex]

Is your feature request related to a problem? Please describe.

As a Thoth operator, I would like to know how many requests made to the system failed on an unsolved package. For this, I would like to observe a metric - the number of advisers that failed due to an unsolved package and correlate it to number of total adviser runs.

Can read as:

X% of advisers failed on unsolved package

This will expose how the data stored in the system are supporting Thoth user needs.

[pacospace]

Is your feature request related to a problem? Please describe.

As a Thoth operator, I would like to know how many requests made to the system failed on an unsolved package. For this, I would like to observe a metric - the number of advisers that failed due to an unsolved package and correlate it to number of total adviser runs.

Can read as:

X% of advisers failed on unsolved package

This will expose how the data stored in the system are supporting Thoth user needs.

We already collect every day this metric in adviser-reporter and store in on Ceph. We could send these metrics to pushgateway to have them in Grafana since there is no ETA on Superset. wdyt?

On Grafana currently we can see the number of unresolved messages sent by adviser workflow and exposed by investigator. This is shown in in the dashboards already: https://grafana-route-opf-monitoring.apps.zero.massopen.cloud/d/6576535a8b26fd1a0e5328f4413dfd8bc989f54f/thoth-kafka-consumers-and-argo-workflows?orgId=1&refresh=30s. This cannot be used for the metric requested in the issue though as adviser workflows are async.

[goern]

/triage accepted
/priority important-soon

[pacospace]

Same as #725 (comment) we can see the percentage of failed workflows and the reason, including if it is due to unresolved which is https://thoth-station.ninja/j/unresolved.html if I'm correct? wdyt @fridex ?

[fridex]

Same as #725 (comment) we can see the percentage of failed workflows and the reason, including if it is due to unresolved which is https://thoth-station.ninja/j/unresolved.html if I'm correct? wdyt @fridex ?

Nice! Then I think this can be considered as done. 👍🏻

/close

[sesheta]

@fridex: Closing this issue.

In response to this:

Same as #725 (comment) we can see the percentage of failed workflows and the reason, including if it is due to unresolved which is https://thoth-station.ninja/j/unresolved.html if I'm correct? wdyt @fridex ?

Nice! Then I think this can be considered as done. 👍🏻

/close

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.