Orange Factor 2: The redesign

A few weeks ago we (jgriffin, mcote, jhammel, ctalbert and harth) had finished most of the bugs and requests that had been made for Orange Factor.  We also did a lot of rework to move it away from couchDB.  Here are some highlights and features that make Orange Factor better.  For reference, here is the original post I did announcing orange factor.

We redesigned it to be simpler and smarter.  First off, the top menu is simpler to use:

top of the index page, menus and important information

You will notice that we default to the last month for the time range.  This can easily be edited by clicking the little calendar icon by the dates (on any page).  Also, we have drop down menus to filter the overall orange factor view based on platform, build type, branch, and test type.  Alternatively, you can enter a bugzilla (quicksearch) query here to see a list of bugs that are related to orange factor.

Next, you can see that our graphs are updated and a lot cleaner:

customizable graph

In this graph, you can see that we have a legend for 4 lines, but only two are showing.  You can toggle the lines on/off.

Lastly on the homepage, you will see a list of bugs related to the graph above.  This is in a new sortable table and we have a lot more data available (some views have more data than others):

list of bugs

These changes are found on all the pages.  For example on the simulator page, we have:

example simulator graph

This page and graph are much easier to read.  We can also send a bugzilla query to see what the orange factor would be if you fixed the query (i.e. bugs assigned to you or your component).  We made some graphs easier to read such as the top bugs graph:

top 10 bugs for orange factor in the last month

In addition to enhancing the website, we added a few integration points with other tools such as tbpl:

link to orange factor from tbpl

Speaking of TBPL, when you star an orange, the data is pushed into our database.  This database is hosted by the metrics team and stores log files and tbpl data.  Actually we pull from that directly and can calculate by test run instead of by raw number of hg pushes.  All of this gives us significantly more accurate data than before with a design that can support other branches.  Another thing we improved was to not store bugzilla titles in the database.  These change over time and waste a lot of space.  Instead we query BZApi to get the title for each bug we are currently displaying.  Another thing we do with bugzilla is harthur added to the bztweaks addon so we can display the orangefactor for a given bug while viewing it in bugzilla. Lastly we have a weekly email that goes to dev.planning which lists the top 10 bugs from the past week and some stats about them.

All in all we fixed about 60 bugs, but feel free to file more.


Filed under Uncategorized

3 responses to “Orange Factor 2: The redesign

  1. Can you detail on the “move away from couchdb” piece?

    Was that moving away from, or moving towards something else, and why?

    • mcote

      I think it was a combination. 🙂 We moved from CouchDB to Elastic Search, for two reasons: speed and flexibility. The former is fairly obvious: ES was much faster than Couch. For the latter, we found ourselves making a series of Couch “views” that were nearly identical, and Elastic Search made it much easier to make those queries more generic and share code.

  2. Hmm, I’m pretty sure that Heather has not added that feature to Bugzilla Tweaks yet. 😉

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s