danisyellis / ddtakehomeosadvocate Goto Github PK

You'll see a lot of metrics because of the DD Agent that I had to install to get the free trial started. You can ignore those.

The number of pull requests open at the time the metric was grabbed.
This metric can be used by project maintainers to, at a glance, see how many PRs are currently open

The number of issues open at the time the metric was grabbed.

• If this were a production project, I'd set it up to run once a day to grab these metrics.

• I would also, of course, remove secrets like the API key from the code and instead have them be environment variables.

• I would paginate the response I'm getting back from the GitHub API to ensure that we get all event data.

• I would add tests. I would mock out the http calls and make sure that calling run in index.js a) gets data from getOpenPulls() and getOpenIssues() b) calls submitMetrics with the correct arguments

• Pull requests opened per day

  • This is the number of PRs opened in the last 24 hours. It would most likely be used on a graph that's averaged, to see the average # of open PRs over time. This could give info on things like project popularity and growth over time.

• Number of pulls by org members/company employees vs. outside contributors

• Number of Active Recurring Contributors

  • Active Recurring Contributors are people who have contributed more than once over a given timeframe- for example, 3 months

• Time till close for both PRs and issues

  • This metric would be useful to look at in a few different ways. Thngs like:
    • a single number of days, averaged, from data from the last 3 months or so
    • a graph that would show any times there were outliers (example: maybe the TTC gets especially big during Hacktoberfest and so more resources should be allocated next October)

• Time till first response for both PRs and issues

  • This would look at the first comment made by someone other than the submitter (or maybe the first comment by someone who's an Maintainer/Admin of the repo )

I completed this project in the allotted 3-4 hours, but I am using a previously created template as a base. The template gave me the skeleton of a project with TypeSctipt, Eslint, and Jest hooked up. (There are similar templates available as Node libraries that I could have pulled in, but I prefer my own.)

The docs for Datadog's Node.js API Client have a word missing that makes the instructions a bit unclear. So I created an issue for it DataDog/datadog-api-client-typescript#1573 (And will potentially make a PR to fix it, depending on the maintainer's response)

• clone this repo locally
• cd into the folder holding the project
• add the github and datadog API keys (found in the notes of my submission on Greenhouse) to datadog.ts line 5 and github.ts line 3
• then run
  • npm install
  • npx tsc
  • npm start