We would like to improve our observability by creating dashboards that monitor CPU / Memory and other metrics across a variety of services. We use New Relic as our Application monitoring service and they have the capability to install a “infrastructure” agent that can report these metrics. Have you seen any successful implementation of this agent deployed with a docker container service? It is typically run as a separate container on the same host which doesn’t seem to be possible with render, and I don’t think we can install on it on the underlying linux instance either.
I saw there was an old example repo with something similar with Datadog that was implemented as a separate Private Service, but im not sure this would work as the agent needs to scrape the information from the host directly.
Since the underlying infrastructure your Render services run on is shared with other customers/services, I don’t believe this New Relic infrastructure agent scraping host metrics would be the right solution for you.
For example, the Private Service running the DataDog Agent example you mentioned involves you writing some code to send the relevant container metrics to the agent. You’d need to be able to take a similar approach with New Relic. Installing an agent like this as a Private service is convenient since you can send metrics from any of your other services running as Native Environments, etc. Let me know if you can set something like this up and I can help you find the relevant metrics.
That being said, this general use case is something Render would like to help solve of the box. It would be great if you could upvote either of these existing feedback items if it sounds like they match your use case. It helps to include as much context as possible about your use case, the problem you’re looking to solve to help us develop the best possible solution.
We rely heavily on customer feedback as a part of our planning and product roadmap process, so capturing interest on the feature request page is very helpful.