What Not To With Mulesoft Composer | What ALL Salesforce Architects Should Know скачать в хорошем качестве

What Not To With Mulesoft Composer | What ALL Salesforce Architects Should Know

What Not To Do On Salesforce: Mulesoft Composer Edition! Things to avoid when using Salesforce’s declarative integration tool. What Not To Do with Mulesoft Composer: 🌎 #1: Data Localization 🌎 🚫 Don’t ignore data localization when designing integrations. 💡 Do understand where your data is being processed in transit. As of 2/23 all Composer resources are provisioned in US or EU regions within AWS. This means that even if the two systems you are connecting via a Composer connector store their data in a specific region (for example, Australia), Composer will run and process that data in a non-ANZ region (like US East). This might be an issue if the data integration you’re designing requires strict regional data residency outside of the USA or EU. 📚 Learn more: https://architect.salesforce.com/well... 📞 #2: API Calls 📞 🚫 Don’t forget that Mulesoft Composer runs on the AnyPoint Platform which is outside the "core" Salesforce platform. As such, API calls can be consumed quicker than you might think when using the Salesforce connector and LDVs. 💡 Do take API limits and data volumes into account when designing Composer flows. If you trigger your Composer flow using connections, one API call will be consumed per record. If you trigger your Composer flow using a scheduler, records will be processed in a collection using a single API call. 📚 Learn more about the importance of bulkification: https://architect.salesforce.com/well... #3 State Management 🚫 Don’t implement Composer without understanding how it manages state. ✅ Do remember that Composer is built on a container-based infrastructure that is designed for resilient and reliable execution of small stateless operations. If there is a failure, the underlying infrastructure (pods, apps, etc.) will auto-recover but the failed process will not auto-restart. Transactions must be re-run or re-triggered in order for data to be processed. 📚 Learn how to manage state here: https://architect.salesforce.com/well... #4: Application Lifecycle Management 🚫 Don’t try to automate your Composer Flow deployment with CI/CD. Composer does not have a sandbox environment. As such, Composer Flows are not supported in metadata deployments and their configuration cannot be exported or managed in source. ✅ Do manually configure your Composer Flow connections via the UI in a single environment. Take advantage of Composer features like flow cloning and be sure to swap your connections from sandbox to production (and vice versa) where necessary. 📚 Learn more about application lifecycle management here: https://architect.salesforce.com/well... ⏰ #5: Real-Time vs. Near Real-Time ⏰ 🚫 Don’t forget to evaluate the latency that your business will tolerate as a part of system performance. When Composer manages events (like changes to a record) it actually polls the system in 15 second increments to pick up those changes. ✅ Do: Evaluate if the business truly needs real-time (and not near real-time). Keep an eye out for (safe-harbor) the ability to configure your own polling increments...it's on the roadmap! 📚Learn more about performance: https://architect.salesforce.com/well... #WNTDOS #WhatNotToDoOnSalesforce #WhatNotToWithComposer Connect with Salesforce Architects: Website: https://architect.salesforce.com/ Twitter: https://twitter.com/SalesforceArchs?s=20 LinkedIn:   / salesforce-architects