1 Some systems create logs continuously, while others produce data only when an unusual event occurs. It’s important for teams to continuously optimize their systems to only collect useful information from logs. Logging levels (warn, fatal, error, etc.) can help you not only filter and extract useful information but avoid information overload. With logging levels, you can monitor some critical events and ignore others.
2 Because most log files contain unstructured text data, analyzing them can be painful. Though modern tools can help you analyze all types of structured and unstructured logs, doing so can take time and is often error prone. It’s easier for to process or parse logs when they’re formatted in a familiar and standard manner. Therefore, you should consider converting your unstructured data into a structured log format such as JSON. When written in a common format, logs help you get faster results for search queries during troubleshooting.
