The solution to this problem is to delegate logging and tracing responsibilities to a separate service which can consolidate diagnostics information from multiple Microservices/Microservice-instances and persist it in a centralized store. This also helps in decoupling the diagnostics activities from business logic there by enabling the ability to scale services independently. The diagnostic service should also be capable of providing a live stream of trace information apart from persisting logs on secondary storage. Centralized logging also enables opportunities for drawing analytics around scenarios like patterns of faults, business trends, and threats. This service can be enhanced by coupling it with notification capabilities to send alerts to the operations team on a specific system state or event. Following is an illustration of a Centralized logging service in action:

Microsoft Application Insights can be used to consolidate logs and tracing information from Microservices deployments. Application insights has the following inbuilt capabilities which simplify overheads around diagnostics for Microservice deployments:

• Centralized logging
• Proactive detection
• Continuous export (to blobs)
• Ability to search and filter logs
• Performance metrics
• Interactive data analytics
• Multilingual SDKs

More about application insights capabilities can be found at: https://azure.microsoft.com/en-us/services/application-insights/.