API Implementation: Using Idempotent Filter to Prevent Message Twinning
Join the DZone community and get the full member experience.Join For Free
[This article was written by Adrian Hsieh.]
Fortunately, the idempotent filter comes standard with Mule ESB, thus greatly simplifying the tasks of preventing message twinning for developers who are implementing REST APIs. The example here will illustrate how the filter can be used in an APIKit project. As illustrated in the application diagram below, first the message must be serialized so that the payload can be properly evaluated for comparison purpose. In this case a Byte-Array-to-String transformer is used immediately after the HTTP listener. Then, an idempotent filter is added in front of the APIkit Router for uniqueness check.
Note the configuration of the idempotent filter. For APIs, the id expression has to be a combination of both the payload as well as the operation being invoked, since the same data can be valid for different operations. Therefore, we used an id expression that concatenates the request path (operation) with the payload.
<idempotent-message-filter idExpression="#[message.inboundProperties['http.request.path']+message.payload]" doc:name="Idempotent Message"> <in-memory-store entryTTL="120000" expirationInterval="1800000"/> </idempotent-message-filter>
Also keep in mind that for the object store configuration, the time-to-live (TTL) parameter unit is in millisecond. TTL value should be set large enough to provide effective filtering but not too large to block legitimate messages from coming through. (For high availability (HA) implementation, choose a cluster-aware object store.)
Twinning is among the more popular activities of millennials hipsters. However, when it comes to API implementation, it can be a unwanted evil that should be exorcised. And for us Gen Xers with the benefits of wisdom from life experience, we also remember that you can have fun twinning when each member of the twin is as unique as it comes…
Published at DZone with permission of Ross Mason, DZone MVB. See the original article here.
Opinions expressed by DZone contributors are their own.
Integrating AWS With Salesforce Using Terraform
Introduction To Git
Scaling Site Reliability Engineering (SRE) Teams the Right Way