APM is about monitoring of application performance, e.g. response time, latency, etc. Metrics typically point to host resource usage, e.g. CPU, memory, network, disk usage. They both give you an overview of how things are going in general in your environment, so they’re both important. But to find the root cause of anomalies, logs are the single common thread. Developers still need to go back to logs to find errors and exceptions when things go wrong. So short answer is that you need both logs and metrics, ideally in a single interface. Loggly is a good option that combines both.