Straight into the service manual page
WrenchQuarry
More

Jira automation hold limits: 2 documented waits for updates

WrenchQuarry Editorial Team · Published · 12 min read · How these guides are written

Two documented Jira automation limits govern a hold: a snooze from 1 minute to 7 days, and a connector resume window up to 15 minutes. Knowing where each limit lives tells you how long the wait can last and what to watch while it waits.

Key takeaways

  • The rule-level snooze accepts a custom duration from 1 minute up to 7 days, and it supports date and time smart values.
  • The connector resume window for AWS SSM and Azure Virtual Machine success responses runs up to 15 minutes.
  • The two limits sit on different layers and do not extend each other.
  • Condition blocks run before actions, so build the branch logic first, then attach a snooze to the branches that need one.
  • A held update that never releases looks the same as one still waiting, so give each hold a named owner and a check-in time.

What a snooze really does to a customer update

When you hold a customer update in a field-service tool, you are borrowing time from someone else's clock. The message does not disappear. It waits. And waiting has its own rules.

Two documented limits set the outer edge of that wait. One comes from the automation rule layer that queues your update. The other comes from the delivery path behind it. If you hold a message past either limit, the update stops being a delay and becomes a decision someone else made for you.

Dispatchers and team leads feel this first. A technician is running late. The customer is standing outside. You tell the system to hold the message for an hour so the crew can get closer. But is an hour inside or outside the allowed window? Does the tool you use actually support that window? And what happens to the update if the connection behind the message breaks while it sits there?

This post walks through both limits in plain words. It covers where each limit lives, how a dispatcher tests it before a real shift, and how to build a hold routine that does not leave a customer in the dark. If the visit itself has to move, see moving a booked visit.

Where the first limit lives: the rule's snooze duration

The first limit is set inside the automation rule that holds your update. In Jira's automation action set, a snoozed action lets you choose a custom duration. The documented range runs from 1 minute to 7 days.

The same action also supports date and time smart values. That means you can anchor the release to a specific moment instead of a fixed offset. For a dispatcher, that is useful. You can hold a customer update until a known appointment time rather than guessing at a number of minutes.

Atlassian documents this behavior on its Jira automation actions page. The page describes the snoozed state as part of the action configuration, alongside other action types. Read it before you build the rule.

Here is what matters operationally: the snooze duration sits inside the rule, not inside the customer's view of the message. That means your hold routine needs its own reminder that the hold exists.

Try this before your next shift. Build a test rule with a 1-minute snooze. Send a held update. Watch the release. Then build the same rule at 7 days and confirm the release does not fire inside your observation window. Both ends of the range deserve a check.

Why the range matters to a small team

A 1-minute floor is useful for sequencing. If two updates would otherwise collide, a short hold lets the first land before the second goes out. It is not a delay for the customer's benefit. It is a spacing tool for your own outgoing stream.

A 7-day ceiling is useful for scheduled work. If you promise a follow-up next week, you can hold the message against that horizon. But a week is a long time for a customer to hear nothing. The limit tells you what the tool permits, not what a customer will tolerate.

The gap between those two ends is where most dispatcher decisions sit. A two-hour hold for a late technician. A four-hour hold for a weather delay. An overnight hold for a next-day slot. All of these fall inside the documented range. None of them are approved by the tool. You are choosing them.

That choice should be written down. Put the durations your team actually uses into a short internal note. When a new dispatcher takes the board, they should not have to guess whether an overnight hold is allowed.

Where the second limit lives: the connection behind the hold

The second limit sits one layer deeper. It is not about your rule. It is about the system the rule waits on.

Jira's automation actions include connector steps that reach out to infrastructure services. Two of those steps have their own resume window. If a flow is set to resume on a success response, it will only resume if that response arrives within a specified duration. The documented maximum for both AWS SSM and Azure Virtual Machine responses is 15 minutes.

Atlassian describes this resume behavior on the same automation actions page. The wording ties the resume to the success response arriving inside the window. A response that arrives after the window does not resume the flow on that attempt.

For a dispatcher, this limit matters when your hold routine depends on an external check. Suppose your rule waits for a status change in a system outside your field-service tool. The rule holds the customer update while that check runs. If the check does not return a success inside the window, the flow does not resume the way you expected.

This is a different kind of waiting. The first limit is a queue. This one is a handshake. You are not holding a message on purpose. You are waiting on a reply that may never come inside the timeframe you assumed.

Test this the same way you test the snooze. Configure a connector step with a short window and confirm the resume behavior. Then configure it at the 15-minute maximum. Watch what happens when the external response is slow. The test tells you how your rule behaves when the network is unkind.

Here is the small comparison, drawn only from the documented facts:

BehaviorDocumented rangeWhat the dispatcher controls
Rule snooze duration1 minute to 7 daysWhen the held update releases
Resume window for AWS SSM and Azure Virtual Machine responsesUp to 15 minutesWhether the flow waits on an external success response

These two ranges are not interchangeable. One is how long you choose to hold a message. The other is how long a connector step will wait for a success response before the flow stops resuming on that attempt.

Holding a customer update without leaving them in silence

A hold is only useful if the customer still knows something is coming. Silence is not the same as a delay. For the wording of those messages, see customer arrival update planning.

Microsoft documents push notifications in its Field Service mobile app for events such as new booking assignments. That is a technician-facing channel. It tells the person in the truck that work has moved. It does not, on its own, tell the customer anything.

Read the distinction carefully. The notification reaches the assigned tech. The customer update is a separate message, often held by a separate rule. If your hold routine assumes the technician will handle the customer conversation, you are routing a customer message through a channel built for the crew.

A cleaner routine separates the two. Let the assignment notification do its job for the technician. Let your held customer update do its job for the person waiting at the address. Then choose a hold duration that respects both.

Some dispatchers hold every customer update for the same length of time. That is simple, and it fails in two directions. Short jobs get a delay they did not need. Long jobs get released before anything meaningful has changed.

A better routine is conditional. Release early when the change is minor. Hold longer when the change is substantial. Jira's automation actions document suggests using If / Else condition blocks earlier in a rule to control when an update is posted. The example given is limiting a post to P1/P2 incidents or to a status change to In Progress. Apply the same idea to customer updates: decide, in the condition block, which changes warrant a hold and which should go out immediately.

Build the condition before you set the duration

The condition block runs before the action.

Write the condition in plain language before you touch the configuration screen. For example: send immediately when the arrival time has not changed; hold when the arrival time has moved; hold longer when the job has been reassigned to a different technician.

Each of those lines becomes a branch in the condition block. The snooze duration then sits on the branch that needs it. The branches that do not need a hold stay quiet and fast.

This approach also gives your team a shared vocabulary. When a new dispatcher asks why a message is sitting, the answer is in the branch, not in someone's memory.

What happens when the hold itself fails

A held update that never releases looks exactly like a held update that is still waiting. You cannot tell them apart by looking at the board.

Stripe documents a similar failure pattern in its invoicing workflow. When a webhook is failing, Stripe sends an email to notify you. That is the kind of signal a hold routine needs: an out-of-band notice that something on the delivery path has stopped working.

Stripe describes this notification on its status transitions and finalization page. The point for a dispatcher is general. If your held update depends on an event that fires when the hold releases, and that event fails to deliver, you need a channel outside the held message itself to learn about it.

Jira's automation actions add one more operational detail worth knowing. After you add a new user to a rule, it can take up to 15 seconds for that user to appear in searches and for the situation to resolve itself. This is a short settling window for a new user, not a hold on a customer message.

Treat that as a short wait after adding a user. Once a new user is added to a rule, confirm that user appears in searches before you rely on the rule for a live customer update.

Give the hold a visible owner

Every held update should have a name attached to the person watching it. Not the person who built the rule. The person on shift who will notice if the release does not fire.

Write the owner on the board next to the held update. When the shift changes, the incoming dispatcher picks up the watch. This is a practice, not a setting.

Pair the owner with a check-in time. If the hold is set for two hours, the owner looks at it before the two hours are up. A hold that is about to release is a good moment to confirm the message still says the right thing.

A shift routine that respects both limits

Here is a routine built only from the documented ranges and the general practice above. Adapt the durations to your own operation.

Start the shift by confirming your rules. Check that the snooze durations on your active rules fall inside the documented 1-minute to 7-day range. Check that any connector step with a resume window is configured within the documented 15-minute maximum.

During the shift, route each new customer update through the condition block first. Decide whether it sends immediately or holds. If it holds, record the owner and the planned release time.

When you add a new user to a rule mid-shift, allow up to 15 seconds for that user to appear in searches.

At the end of the shift, hand over the list of held updates, using the Field Service Work Order Checklist for Dispatch Handoffs. Each entry needs an owner, a release time, and a one-line reason for the hold.

This routine uses the limits as boundaries, not as defaults. The documented ranges tell you what the tool accepts. Your operation decides what it uses inside those ranges.

Check this before your next held update

Use this list on a quiet moment, not during a live hold.

  • Confirm the snooze duration on the rule sits inside the documented 1-minute to 7-day range.
  • Confirm any connector step with a resume window is set within the documented 15-minute maximum.
  • Confirm the condition block runs before the action, so only the intended branches hold.
  • Confirm the customer-facing message and the technician assignment notification are separate.
  • Confirm the owner and release time are written down for every active hold.
  • Confirm any new user added to a rule mid-shift has had up to 15 seconds to appear in searches.

Your next step today: open one active rule and confirm its snooze duration sits inside the 1-minute to 7-day range.

Related guides: Florida contractor delays: 2 timing windows to check · New entrant audit timing: 2 clocks to track.

FAQ

What is the shortest hold the tool accepts? One minute. The documented floor for a custom snooze duration is 1 minute. Anything shorter is outside the accepted range.

What is the longest hold the tool accepts? Seven days. The documented ceiling for a custom snooze duration is 7 days. The same action also supports date and time smart values, so you can anchor the release to a specific moment inside that window.

How is the connector resume window different from the snooze duration? The snooze duration is how long you choose to hold a message. The resume window is how long a connector step will wait for a success response from AWS SSM or Azure Virtual Machine before the flow stops resuming on that attempt. The documented maximum for that window is 15 minutes.

What should I do when a held update does not release on time? Check the rule's snooze duration and condition block before the connector step. If your delivery path can send an out-of-band notice on failure, use it. Stripe documents this pattern on its invoicing webhook failures, where an email notifies you that the webhook is failing.

Who should I ask when a rule's behavior is not covered by the documentation? Ask your tool's support team for the specific action you configured. The pages cited here document the ranges and the resume behavior. They do not cover every combination of conditions your operation may run.

Does the technician push notification replace a customer update? No. The Field Service mobile app documents push notifications for events such as new booking assignments. That reaches the assigned technician. The customer update is a separate message with its own hold. Route them separately.

Sources

More field guides