The points listed here can be considered as an advisory note:

• Enterprise tracking and locking mechanisms are deployed as a centrally accessible service, which does not have to be implemented as a central logger, but mainly as a distributed logger and message tracker.
• Logging is usually implemented asynchronously. This means that all the business SOA services perform the logging part by invoking log-related services asynchronously, and usually, via the SOA services library.
• Tracking, logging, and error logging mechanisms are often implemented as a combination using the same network traffic.
• Since the traffic which has the highest frequency in the whole SOA platform is the tracking and logging traffic, it has to be implemented with the utmost care with every minute detail. It is also the most critical of all other types of services in the SOA platform.
• It is not necessary to implement all logging via the same mechanism or same set of services. For example, we can have a central logger for certain log levels, and perhaps, logstasher components (per machine) with ElasticSearch database for another set of log levels.