Open asajeffrey opened 5 years ago
@asajeffrey I would be happy to help with this. Honestly a little new to d3 but willing to learn :)
Excellent, hopefully this won't be too too complex. There are three dashboards at https://github.com/servo/servo.org/blob/gh-pages/dashboards/index.html, do you want to pick one of them and look at the raw data on s3? They're all pretty straightforward I hope!
Rather annoyingly, https://github.com/servo/servo/issues/22462 means we're missing a lot of performance data, you'll need to go back to July 2018 to see data.
@asajeffrey Thank you very much for your patience, I apologize for the delay in progress! I plan to make steady progress on this in the coming weeks.
I am having trouble understanding how to read the raw data from s3. I described my attempt below, I would appreciate some guidance in accessing the raw data.
I navigated to https://servo-perf.s3.amazonaws.com to see the raw data but it's unclear to me how to access it. I tried parsing what is shown at the address to form a link to a CSV (I recall I could produce a CSV from s3 in January/February) but now it returns NoSuchKey
for links /servo-perf1000falsebuilds-2016-10.csv2018-12-13T02:44:22.000Z"e6468dd9cd104cf86c9def9ef852bb15"2227319f573799c220ae89a89716cb0d49f3e0319ff06706917404dd5bdef5eb2a90b07mozilla-servo-awsSTANDARDbuilds-
(from a line entry), /buildTime
(a key I thought might be a valid key), /2019-09-15
(a date I though might be a valid key)
For each <key>
, you can append that value to the URL, like https://servo-perf.s3.amazonaws.com/builds-2016-10.csv.
We were using Google Data Studio to generate reports on things like servo build times, but we were on a free plan, and Google moved our reports to the trash. They were quite simple graphs built from CSV files, I suspect they wouldn't be too difficult to replicate in something like d3. Anyone willing to give it a shot?