Typical bandwidth usage
How often does Netdata report data to the cloud dashboard?
How much bandwidth usage is expected per node?
cjcox last edited by cjcox
My understanding it that it works similarly to how graphs are done in Netdata client. So, time data can be “viewed” and the range data applicable send to the cloud, and real time, where the “ticks” are sent to the cloud. In many ways, the cloud is more similar to how netdata client in the browser works.
So, expect the same kind of traffic (maybe less) that you get from a browser to a netdata client. Realizing that you could have many netdata client hosts reporting data to the cloud, but afaik, like the normal browser to client, only data going is for what is showing. There’s obviously a population send for “the view” at first graph view of a graph.
I supposed it’s possible (maybe) to create a dashboard or even war room from hell what shows many graphs on a really big/wide browser display that could be shocking…
I’m not sure if they are trying to do any caching cloud side. In fact, we’ve been told that data sent is not stored. Just some of the meta for origanizational reasons (e.g. node names, element names).
cjcox last edited by
Why caching might be interesting… in the case of multiple people viewing the same graphs. So… they might be doing that (??)
When agent connects we send some metadata (chart list etc.).
Actual data are not stored (on the cloud) and are sent only when you are viewing the charts that need them on the Cloud dashboard.
Some small things like alarms are pushed by the agent when needed but those messages are small. Also there are keep alives but those are tiny.
As for caching someone from the cloud team will have to answer that.