Edge HA configuration

In the agent config file you can specify more than one Server/Edge using a comma separated list but we were asked if there was an option to configure the Edges in an HA configuration. What we would like to know is what are the options for resiliency in the event that an Edge is unavailable to allow continued attribute collections? 

  • We do currently have the ability to set up an active/passive type edge HA scenario, but it is quite manually configured.
    However, we have a near term roadmap task that is currently in development known as our "Clustered Edge" which means you'll only ever need 1 "Edge device" configured in your Stratis/DDM per datacenter. This means that there will be a cluster of edge devices that all will in concert to perform the tasks of many edges, which also handles the redundancy challenges.
    This also makes it much easier to deploy agents, as they will all use the same self administrated VIP address to interact with agents and incoming netflow/sflow traffic, etc.
    One cluster could be 3 VMs working together, or it could be 50+ depending on the workload.
  • In reply to bryancan:

    Thanks Bryan, do you have more information you could share on dates for the availability of 'Clustered Edge'? Knowing how soon that should be available may shape whether we would need to pursue the active/passive config.
  • In reply to drafn:

    The clustered edge release candidate was made available by invitation in the past week. It is a retrofit (is added to Stratis 3.2.7 for example). I suggest you request the FCE doc for more information.