Best practices for optimizing automation flows

The following best practice guidelines will allow you to optimize your automation flows, so that they run more smoothly and are easier to manage.

Manageability

Plan your flows before creating them

Planning your automation flows ahead of time will help you work out what flows you must have, and where you can instead combine flows by using conditions.

Start with your team’s work processes and the actions you want to occur, then work backwards to what events you want to trigger those actions. For example do you want to send emails in response to customer feedback comments on all work items, or just for high priority work items?

Align your automation strategy with your business and operational processes

Automation flows are designed to help your business and operational process work more efficiently; they should align. A review of your automation flows should occur alongside reviews of your business and operational processes.

Staying within usage limits

Add conditions to flows

Use conditions to ensure that actions occur exactly when you want them to. Only flows that perform a successful action will count toward your usage, so this will help you make the most of your monthly limit. For example, in Jira automation, consider adding a filter for work item status, to only run for particular statuses (like To Do, or In Progress), 

Check the scope of your flows

Make sure your flows run only in the Jira or Confluence spaces you want them to run in. That is, if you have flows with Global scope but they don’t need to run across every Jira or Confluence space, change their scope to multi-space.

This will minimize wasteful flow runs, and help you get the most out of your monthly usage.

Disable flows

Monitor your flows, and identify the ones that are either running too much, or flows that you don’t need, and disable them to prevent wasteful flow runs.

Performance

Restrict the scope of your flows

Avoid using global scope as much as possible. Consider whether flows must apply to multiple spaces or if a single-space scope would suffice. This change improves the efficiency of your flows: any events from spaces without a matching flow are filtered out. As a result, there are fewer flows that need to be queued for execution.

Combine flows

Each flow may function perfectly well alone, but flows can often work better together. Are you able to use the same flow for multiple current flows? For example, you might have a flow that branches on the current flow and then runs conditions and actions underneath.

Select an appropriate trigger

Some triggers are more specific than others, so if your situation requires a specific trigger, avoid using a more general one. For example, in Jira automation, Field Value Changed is far more economical than Work Item Updated. This is because updating a work item can encompass many things, while changing a field value is a more specific kind of update.

Use Compare conditions as early as possible in the flow chain

You want to use conditions that compare one value with another early in your flow, to eliminate possible work items or pages as early. Examples of compare conditions include the User condition, Advanced compare condition, or Work item fields condition (Jira only).

For example, if you used a condition early in your flow to check if work item type = Bug, this makes your flow much more efficient because you’re eliminating all non-bugs from your flow early.

Ideally, you should be chaining these conditions at the start of your flow to exclude all the work items you don’t want to operate on as early as possible. 

Move checks that require more processing further down the flow chain

Conditions that check singular fields, such as the Work item fields condition (Jira only) or User condition, are the simplest to run so these should be earlier in your flow chain.

The Advanced compare condition, and those involving query languages like the JQL and CQL conditions, require more processing. Try to avoid these if possible, or push them to the back of the chain.

For example, checking if type = Done (a JQL condition) requires more processing than comparing {{issue.status.name}} equals Done (an Advanced compare condition).

Use branches sparingly

Only add branches to your flow where you absolutely must. This is because branches spawn new processes in your flows. Branches are useful if you’re combining flows, but if it doesn’t serve a purpose then it can have negative effects on performance.