I've installed and configured apm-server in an EC2-instance in AWS. I've installed RUM agent on a ReactJS application as well as NodeJS apm agent on my Nodejs backend api server. We don't want to lose any logs. I noticed on the frontend application, it makes this POST call, /intake/v2/rum/events. What can we do to make sure the events being sent by RUM agent or even the node-apm agent doesn't lose the logs in case our apm-server cluster was unreachable? I'd like to buffer all the logs when our in-house apm-server is unreachable or down. Then resume shipping it when apm-server becomes online.
@Nie https://www.elastic.co/guide/en/apm/server/7.3/common-problems.html#server-es-down describes what happens when the APM Server is unavailable. Some of the agents buffer a limited amount of data, but the Node.js and RUM agents do not, and there is presently no way to change that behaviour.
Even when the APM Server is available you should not expect agents to capture 100% of transactions in high-throughput applications, as that may impact application behaviour.
Can you please explain why it is important to you to guarantee delivery?
Sorry took me a while to reply. Actually, I forgot about Filebeat. My question was really related to Filebeat since it ships our Nginx logs. APM is meant for monitoring. We don't want to lose any Filebeat logs for the purpose of troubleshooting issues. I think we are retaining 7 days worth of logs then we dump it. I'll have to double check on that.
Btw, that is a very helpful link. Thank you.
This topic was automatically closed 20 days after the last reply. New replies are no longer allowed.