When building distributed applications we need to make sure that when we talk to a remote services or our microservices talk to each other we can handle downstream call failures and either recover automatically or gracefully degrade our own service instead of failing outright. There are two types of downstream failures. Those that are transient in nature (such as network »
Applications running in the cloud or just making use of various external services are likely to face temporary failures when trying to communicate with these services. These failures could be network connectivity issues or request timeouts when external services are overloaded. An important aspect of these failures is that they are self-healing and if we retry calling services after a »
Configuration management is an important part of running and evolving public facing services when down time is not tolerated and you need maximum flexibility in adjusting settings of running services. Implementing External Configuration Store Pattern opens up some great possibilities for you like independent deployment of your services and configuration, sharing configuration between multiple services and versioning of configuration sets. »
In my previous post I provided a sample implementation of the Priority Queue pattern that was multiplexing messages from several queues through a dispatcher to a limited number of worker threads. While the basic idea looks good there were a couple of issues that popped up while running a similar solution in the real environment: Some messages got put in »
One great feature that Semantic Logging Application Block (SLAB) provides is an ability to use out-of-process service to listen to your messages. The idea is to move the logging sinks to another process so that messages don’t get lost if your application process crashes or it is terminated unexpectedly. We use SLAB out-of-process service in our cloud services that »
Last time I was talking about what a priority queue is and how we can approach designing it. In this post I want to guide you through a sample implementation of such a queue that won’t require dedicated computing resources per each priority. The queue is going to be based on Microsoft Azure Service Bus but the approach can »
Queues allow you to decouple components so that they process data or perform tasks in asynchronous manner. This greatly improves scalability and responsiveness of your application. Priority queues allow publishers influence the sequence in which requests are processed. It allows them to shift from a regular FIFO pattern to the one that takes a request priority into account and lets »
I and my team have been using Semantic Logging Application Block (hereinafter SLAB) on our project where we create cloud services running on Azure. It’s important for publicly faced services to log everything so it is possible when we analyze that data to have a decent idea of what happened and how our services were used. You can read »