A webservice for displaying stats from graphite or lgTask.

Requires jsProject: https://github.com/sellerengine/jsProject

pip install cherrypy

Checkout jsProject and lamegame_statview:

git clone git://github.com/sellerengine/jsProject.git
git clone git://github.com/wwoods/lamegame_statview.git

Configure lamegame_statview:

cd lamegame_statview
vim app_local.ini

[source]
type = "graphite"
url = "http://path/to/graphite/web/interface"
authKey = "Basic HTTP_AUTH_CODE" (optional)

python run.py

Point your browser at http://127.0.0.1:8080, and you should see the interface. Before making graphs you'll need to configure paths - part of the power of using the StatView interface is that it's much more intelligent about your paths than Graphite's built in interface is.

• Click on the "Add/Edit Stats" button in the upper right corner
• Click "Add new"
• Enter a path matching your statistics... for instance, if your host load data is stored at hosts.myMachine.load, enter: "hosts.{host}.load", and set the type to "Sample".
  • Counter - Aggregation is a counter - that is, each number is an incrementing event, like the number of times a page loads.
  • Sample - Aggregation is a sample - that is, each number from graphite is a sampled value, like machine load.
  • Sampled max - Different stats are aggregated according to the highest absolute value within all matching stats, rather than adding. We use this for "Max age of task on a given host", for instance. Sample, in contrast, would add all of the max ages for each task type, rather than returning the overall max.
  • You can use "" for path wildcards.. if we had other machine stats, using hosts.{host}. would load all available statistics at that level. "**" can be used to match any depth element.
• Click off the dialog to save
• Click the green "Add New..." to add a graph
• Click on a property to add it to the expression... expressions are simple math equations. You can also use min() and max() easily. See "Expressions" below.
• Click on a group name to divide the graphed data according to an element in the path.
• Click off to save, and look at your new graph! Oo, ahh. Mouse over for tooltips.

That's the gist of it... I'd be inclined to write more docs with more interest.

Thanks for your time!

Click on one to show it. Click and hold for one second to show all.

You can also use "for each {group} add {expression}" for fancy bits. For instance, we have a section that is "Failing expectations." For this, we do something like:

for each user add 100 - actionsCompleted

And smooth that over an hour. The "for each" clause ensures that the 100 is applied once per user, and not once to the sum of all actions completed by all users.

min(a, b) and max(a, b) are also provided.

• 2014-5-19 - fixing an issue with combining two stats that are associated with different sets of groups.