We’re assessing the possibility of running Graylog on AWS ECS and were interested in discovering other people’s experience running the same setup.
There are guides on running Graylog on Docker which we have trialled but we were interested in knowing whether a containerised implementation is truly “production grade” from those who have run containers in a production grade environment.
For those running Graylog in containerised production environments, more specifically AWS ECS, what are the benefits and what are the disadvantages.
If your team is confident administrating a containerised setup via ECS then there will be no real difference between this and a host based cluster. In fact there is some benefit in being able to better scale the cluster based on ingest.
Did you have any specific reservations on this setup?
No, not really. We run other workloads in ECS but have always run Graylog on it’s own dedicated EC2. The additional benefits we see are that we can upgrade Graylog versions easier and you remove the need to worry about upgrading the underlying OS.
Given the continuous ingestion of data into the data-node, we are thinking of running the Graylog cluster backed with dedicated, managed, EC2 instances as opposed to using Fargate.
