September 10, 2019
Sooner or later the system will fail. Doesn't matter if it'll be for an infrastructure failure or a bug. It'll fail. Dealing with failures can be tricky especially when business failures are handled like if they were infrastructure ones.
NServiceBus has a configurable routing engine whose role is to define where messages should be routed when sent. The routing engine is configured along with the endpoint, that might not be the most comfortable solution. Is there anything we can do about it?
When designing UIs for a distributed system we're faced with many choices. It's hard to districate ourselves into the options jungle. We might be tempted to select what at a first look seems to be most flexible solution, as flexibility pretend to imply freedom. It's tremendously easy to transition from freedom to anarchy.
There is search and then there is search. And despite having the same name their meaning, and thus their technical implications, is shaped by the business we are in. Let's have a look at what search in a distributed system is and what problems brings to the table.
HTTP is the Microservices way, so it must be used as a communication transport, they say. Is HTTP adding any benefit to our systems? Can it be considered harmful in some scenarios? Are there better approaches? Shedding some light on the system logical architecture might be helpful.