What makes combining monitoring and log analysis powerful?

The primary reason is that it's useful to know what was happening on a system when something was logged. When troubleshooting an unexpected error in a log, the troubleshooting starts by building context. What was the action, what was the environment, etc. It's often rather difficult to diagnose an error message by simply looking at a log file.

For example, if your API server talked to Postgres and you start to see some timeouts, it's great to open a monitoring view for the same time period and see the overall traffic, the number of connections to postgres and what else was happening during this time.

It’s helpful to be able to correlate the metrics that you’re monitoring with your logs. Once you can connect them, you can really see where the issues are and drill down into the logs to figure out what’s going on. We use Loggly and we’ve found that it saves us significant time.