We chose an AWS t3a.large instance, which is a woefully underpowered machine for most FileMaker or Claris deployments. We have a tool named Punisher that we use to test FileMaker Server it is ideally suited to test the impact of logging since it goes through a series of taxing operations that cover most anything that developers do in their solutions.įor a server, we wanted to pick something low-performing and easily overwhelmed, which should amplify the impact of the extra load added by the logging. So let me put my money where my mouth is. To put it bluntly: if turning on the logging makes the server fall over, then the deployment has much bigger problems than dealing with the little bit of load that the logging adds. When we ask for the logs to be turned on, we must explain that the warning they were so afraid of is meaningless.
Those invaluable logs are turned off, and there is no relevant set of data on which to base any conclusions. When we are tasked to help find out why a particular deployment has performance issues, invariably, we find that there is nothing to go by. That warning is sufficient to scare most customers into hastily turning it back off.Īnd yes, I have a big opinion on this – that warning does more harm than good.
0 kommentar(er)
