A summer 2011 Cloud Foundry Fling
plotscope requires node.js and MongoDB. You can
vmc push to Cloud Foundry
(e.g. a micro cloud running on your machine)
or run on your machine. Here's how to grab the dependencies for local use, as of 9/8/11.
Generally, the downloads page is helpful.
If you're on Debian or Ubuntu, I'd recommend using packages from the apt repos
After the package is installed,
mongod should start automatically.
If it doesn't, run
mongod; the db shell is accessible as
mongo for testing.
The installation page is helpful here. Beware that node is very in-flux, and building from a bleeding-edge version will probably cause sadness. Plotscope has been tested on v0.10.12.
Just do a
from this directory and the app should fire up on
To add data, go to the upload page:
<name> is the name of the dataset you'd like to add and
a scaling value for the data. The axes are currently hardcoded to have a maximum of
YMAX = 1000000; thus, if you have data between 0 and 1000,
<scale>=1000 is a good choice.
The original data values/range will still be maintained, and the UI shows the actual scale on
multiple axes to the left of the plot.
One of the tiny sample data sets is in
etc/data/stock-vmw.bin for testing this.
After uploading, to build aggregates, go to
<name> was the name used for the upload. This creates the scaled views necessary for the UI.
Data is in a simple binary format of 8 byte datapoints: 4 byte timestamps and 4 byte values (little-endian) with no header. Datafiles should look like
Some data sources of note:
- A little synthetic data creator is in
- To convert from the format stored by the SS machines at the PDL,
etc/data/ssmachine.pycleans them up and will spit out binary
- To convert from CSV or delimited format, there's
Surely there are bugs in both what's written above and in the app! Contact email@example.com if/when "stuff breaks" :)