In T364901, T367870, and T367776, we implemented the logging needed to monitor Citoid performance with the intention of:

1. Identifying and addressing issues as they emerge
2. Ensuring Citoid is behaving in ways that meet volunteers and partner needs
3. Evaluating the impact of changes we're making to Citoid evaluating the rates at which domains are failing and succeeding.

This task involves the work of investigating the data that is now available in this dashboard and potentially, prioritizing additional analyses, so that we can answer questions like those listed in the Research questions section below.

Research questions

• What is the overall rate at which Citoid requests are failing?
• Of the Citoid requests that are failing, what percentage originate on/from Wikipedia/the visual editor?
• What domains are failing most frequently?
• Which – if any – third-party “actors” (broadly defined) are responsible for generating outsized portions of Citoid requests?
• What impact – if any – have changes like T367452 had on Citoid outbound request volume and the rate at which those requests fail? [i]
• What percentage of Citoid requests originate on/from Wikipedia/the visual editor? What percentage of Citoid requests originate on/from third-party sites?

Decision(s) to be made

The questions listed above are meant to help us decide the following:

1. What partners should we prioritize opening up conversations with?
2. What architectural changes should we prioritize making?
3. What additional investigations should we prioritize work on?

Open questions

• [ ] 1. What – if any – additional instrumentation do we need to add in order to answer questions documented in the Research questions section above.
  • This question will now be answered in the newly-filed T369663

Observations

This section will eventually contain what notable observations we make through reviewing the data available in this dashboard.

• Slack thread: 26 June
• Slack thread: 28 June

i. Thank you, @Sj for raising this question.