I am new to elasticsearch and have been trying to introduce in my company. Many colleagues have asked a question: how much bandwidth is required and, due to this, our production environment can impact. Please advise how much bandwidth we use
There's a number of consumers of bandwidth in an Elasticsearch cluster. The main two in normal operation will be indexing and search traffic: indexing is putting data into the cluster, and search is getting it out again. These can consume as much or as little bandwidth as you want, they're completely driven by external traffic.
Recoveries also consume bandwidth. Normally there are no ongoing recoveries, but if you suffer some kind of failure then you will need to use recovery bandwidth to build new replicas to replace the lost ones. This is by default capped to 40MB/s per node but you can adjust this limit.
Another bandwidth user is snapshotting (i.e. on-demand incremental backups). This is also capped to 40MB/s per node by default, and this is configurable.
Thanks for the reply. I would appreciate if you can provide any reference where I can get the complete details.
I need to cap the bandwidth.
I don't know that there is such a reference, and there is no overall bandwidth cap within Elasticsearch. To be honest I don't know of any similar systems that provide this information either. Can you point me to some examples of the kind of thing you're looking for?
Of course you can always use an external tool (e.g. a traffic shaper) to limit the bandwidth of any software. Maybe this is what you are looking for?