Administration

The Administration tab in the Concerto panel contains several sections that can be helpful in monitoring aspects of your Concerto environment. These are explained below.

Settings

• API enabled: set whether or not to allow API clients to make requests to this instance of Concerto
• Session limit: set a hard upper limit for the number of simultaneous sessions permitted on this instance of Concerto. If you have a single instance or otherwise non-redundant server set up, this setting can be useful to avoid out-of-memory situations. When the limit is reached, it will prevent further sessions and show a message to new users that there is too much traffic and they should try again later. What value to set this limit at depends on the specifications of the server on which you are running this instance of Concerto. Refer to the deployment guide for more info.
• Session runner service: set whether to use serialised session (default) or persistent session handling for this instance of Concerto. Serialised session is recommended because it has significantly lower RAM usage, which normally means you'll be able to handle a higher number of simultaneous users on an instance with fixed memory capacity. Certain users may want to use persistent sessions in order to bypass the flowchart builder and write tests with pure R code without reliance on the nodes, whereby the user session is kept open in R (which is less efficient than serialisation).

Content

This tab gives you access to manage how you import and export content to and from your Concerto instance. Here you can enable Git integration, allowing you to connect your instance to external repositories and run continuous integration and deployment processes for your projects. You can also import content from a URL, pointing to a single file export (only HTTP/HTTPS protocols allowed) or to a file path. Or you can upload a file from which to import content.

Messages

Here you can see error messages from your instance. You can also see these errors, when they occur, in your browser console while running your Concerto test in debug mode. See here for more info on how to check for bugs.

Usage charts

Here you can view statistics on the number of test sessions by specific date, which can be useful for monitoring usage of your instance and informing whether you need to scale server capacity up or down or modify the session limit.

R packages

Concerto nodes already use features from popular R packages. However in some cases you may want to install specific R packages into your instance, so that you can call them from R. You can do this here, though this feature is somewhat experimental. Note that compiling packages is a resource intensive operation and it is very inefficient to do so through R rather than through a repository with binaries. Some packages may therefore fail to compile if your server isn't powerful enough. For that reason, the recommended path to installing specific packages in your instance is instead to create a custom Docker image with these packages installed in it, and then to host this custom docker image. For example, a Dockerfile to build a custom docker image with the ggplot2 package installed would look something like this:

FROM campsych/concerto-platform:5.0 RUN Rscript -e "install.packages('ggplot2')"

If compilation fails for lack of resources, you may also consider temporarily scaling up your instance, compiling the package, and then downgrading it again.

API clients

Here you can configure and manage clients that make requests to your Concerto instance via the Concerto API. 'API Enabled' needs to be checked in Settings for this to function.