By Boomi
Do you ever wish you could make your Boomi processes run faster?
Integration Consultant Adam Bedenbaugh spends his days working with the Boomi AtomSphere Platform. He knows a lot of optimization tricks — and the most common places to look for flaws that slow down process execution.
You’ll learn how to:
- Use flow control (and what not to do with it)
- Limit API calls to external systems
- Streamline data access with document caches
- Pinpoint and solve recurring problems by going over execution logs
- Determine how long each process component takes by reviewing the process state
To give you a taste of Adam’s session and what you’ll see in his demo, we’ll take a look at his approach to flow control.
The Do’s and Don’ts of Flow Control
Flow control is something everyone likes to use, but there are good ways and bad ways to use it. In this session, Adam walks through how to best set up Flow control, and explains configuration options and the pros and cons of each approach.
There are two configuration options, parallel and batch.
Parallel processing has two flavors — threading and processing. The processing option can be beneficial if you’re running a lot of data and you need more memory. But if you’re trying to increase concurrency, threading may be the way to go.
The batch option has three parts: no document batching, run each document individually (not recommended for most use cases), and run # documents per batch. If you’re looking to control the number of API calls to stay under a specified limit, that’s what you’ll want to use.
Troubleshooting: What to Do When You’re Not Sure What to Do
Let’s say you’ve already done a lot to optimize your processes. Is there more you can do? And what is it? Adam’s hint: review the execution logs. If there’s a recurring error or something just doesn’t make sense, the answer is probably in the logs.
Suppose, for example, you keep hitting an API limit, but you don’t know why. When you review the logs, you might find you have a Try/Catch shape that retries documents. When the first API call fails after you hit the limit, the Try/Catch shape just keeps trying that document. It adds minutes to your processing time, as it stops your process in its tracks.
Second, review process states. The process state can tell you how long each process component takes. In the example above, the process state reveals how long each Try/Catch attempt took.
The logs and process state will tell you what things interfere the most with optimization and where to focus your efforts.