How are Security Gateway availability problems handled?

In general, a Security Gateway is preferred if it is considered to be "closer" to the sender (i.e., OneAgent or another Security Gateway) in terms of this hierarchy: Dynatrace Server < Public Managed Security Gateway < Private Security Gateway.

Entity Priority Description
Dynatrace Server Low This is the default setup and fallback option. Dynatrace server is considered to be "far away" from OneAgents and other Security Gateways.
Public Managed Security Gateway Medium These are typically used as an entry point because the cluster (and its public Security Gateways) aren't directly reachable. Therefore, this is preferred over Dynatrace Server.
Private Security Gateway High The idea behind a private Security Gateway is to install it as close to OneAgents as possible. Therefore, it is considered to be the "closest" Security Gateway.

Security Gateways of higher priority can generally send data to Security Gateways that have lower priority, but not vice versa. Security Gateways can't send data to other Security Gateways of the same priority.

If a Security Gateway isn't reachable, after about 30 seconds OneAgents will try to contact other Security Gateways. OneAgents will always prefer Security Gateways of the highest available priority as long as at least one of them is reachable. If more than one such Security Gateway is available, OneAgents will try to switch between the available Security Gateways on a regular basis to achieve proper load balancing.

If all Security Gateways of the highest priority are temporarily unreachable, OneAgents will switch to the next lower priority Security Gateways and continue checking in the background for availability of higher priority Security Gateways.