Design – Build – Run: Applied Practices and Principles for Production-Ready Software Development

20.4. Summary

This chapter has looked at two important factors for successful monitoring of a solution, performance counters, and events. To support better monitoring, the solution should support a level of configuration.

The following are the key points to take away from this chapter:

• Identify and use performance counters. Walk through the solution components and highlight areas where performance counters will provide real monitoring benefit. What may seem like something irrelevant may have real benefit in production running.

• Updating performance counters mustn't affect performance. Ensure that performance counters are updated only when they need to be and that everyone understands the counter, its sampling rate, and values.

• Think about the events that the system should raise. Not all events are "bad." There are some very good reasons for raising good events. You should walk through the solution and identify events that should be raised to assist in monitoring and operations.

• Use application specific event logs. Determine where events will be raised and the logs they will be raised in. You should consider using multiple event logs where necessary.

• Include detailed contextual information in events. You need to ensure that all events contain all the information they need for the follow-up incident investigation.

• Consider a flexible monitoring configuration. Remember that not all environments are the same. What might be a warning in one environment may be an error in another. Consider ...