When you’re in the middle of a production down event and your whole team is diagnosing the issue, and your log server is unresponsive, who do you contact for support?
No one, you pull an engineer off the production issue to debug the log server, because you need the log server to debug the production servers.
See the problem?
Edit: to be clear I’m no fan of Datadog and I wish self hosting were an option. I want this path for our company, but at least on our team we just don’t have enough (redundant) expertise to deploy and manage these systems. We’d have to hire an extra FTE.
No one, you pull an engineer off the production issue to debug the log server, because you need the log server to debug the production servers.
See the problem?
Edit: to be clear I’m no fan of Datadog and I wish self hosting were an option. I want this path for our company, but at least on our team we just don’t have enough (redundant) expertise to deploy and manage these systems. We’d have to hire an extra FTE.