URL View of Graph Displays Single Point

[rdfedor]

Hello, great product but may have stumbled across an issue. I did a straight clone of the repo and set up the app on my service, and I have several domains, all with the same app, feeding mongodb. I changed the header.php file in the externals folder so that URL gets the $_SERVER['SERVER_NAME'] along with the $uri instead of just the $uri.

Well, the data is being populated as expected, however the problem I'm running into is that the charts on the url view are not rendering properly. I get a chart with 2 points on the far left and that's it.

The JS data that the code is rendering to the page to graph looks like this

var graphData = [{"row_count":6,"raw_index":5.4,"date":"2014-01-17","wt":237285,"cpu":36003,"mu":2259360,"pmu":4168944}];

and the result from the mongodb aggregate function within getPercentileForUrl returns,

Array
(
    [result] => Array
        (
            [0] => Array
                (
                    [_id] => 2014-01-17
                    [row_count] => 6
                    [wall_times] => Array
                        (
                            [0] => 223810
                            [1] => 237285
                            [2] => 191848
                            [3] => 233181
                            [4] => 228890
                            [5] => 243629
                        )

                    [cpu_times] => Array
                        (
                            [0] => 36002
                            [1] => 36002
                            [2] => 28001
                            [3] => 36003
                            [4] => 36002
                            [5] => 44002
                        )

                    [mu_times] => Array
                        (
                            [0] => 2298520
                            [1] => 2259360
                            [2] => 2256872
                            [3] => 2255672
                            [4] => 2179600
                            [5] => 2227840
                        )

                    [pmu_times] => Array
                        (
                            [0] => 4168896
                            [1] => 4168944
                            [2] => 4168896
                            [3] => 4168896
                            [4] => 4168944
                            [5] => 4166640
                        )

                    [raw_index] => 5.4
                )

        )

    [ok] => 1
)

As a side note, I also had to comment out the ENV variable because I was getting errors from mongo stating unable to store empty value error consistently from the header.php file.

Thanks

[markstory]

I know that the URL view graphs have never handled a single point of data well. I should really fix it, thanks for reporting the issue.

[rdfedor]

Well, the problem isn't necessarily that, I have a 6 for the row_count, however it treats it as though its 1 result. I was also getting this error when trying to store the $_ENV data,

[Fri Jan 17 11:12:04 2014] [error] [client 192.168.1.44] xhgui - zero-length keys are not allowed, did you use $ with double quotes?

then after allowing zerolength keys in php, i got this error.

[Fri Jan 17 11:32:26 2014] [error] [client 192.168.1.44] xhgui - '\0' not allowed in key: \0...

After commenting out 'env' => $_ENV, everything started working.

If it helps at all, I used Symfony autoloaders in my application.

[rdfedor]

It looks like the issue was I needed to change the search criteria to the last 30 minutes.

[markstory]

By default the chart defaults to day view, and groups results by day. You can modify the range below a day and the results will not be grouped anymore.

[markstory]

The URL view now has the ability to get smaller time frames and ungrouped results. Does that help solve this issue.

[aheinz-sg]

I had the same confusion. Perhaps adding a label to the x-axis would make the grouping more obvious?

Even when I do a search for last 24 hours, the first and last data points are not marked, which is confusing for small data sets.