Browsium Client is designed to intelligently manage network load to avoid spikes in case many client machines are brought online simultaneously, or in case the Server is temporarily unavailable and then restored.
Proton Controller (component in Browsium Client) instances have independent randomized delays so that when many of them start simultaneously they will not all try to contact the Server at once. In case the Server becomes unavailable or overloaded the Controller does exponential back-off, doubling the time between each communication attempt and thereby reducing the network load. Also, the default time intervals installed with the Controller are relatively long such that large installations will not swamp the network before the Controller instances can even get updated intervals from the Server. Finally, under a prolonged failure to contact the Server, the Controller will stop collecting new data entirely and wait until communication is restored.
Controller stops data collection for any of the following reasons:
Posted in: Proton Knowledge Base