InnerWeb
Skip to main content

Time trigger checklist for notifications: window, sent marker and network failure

A task that sends a notification N minutes after a phase starts looks simple, until it has to be maintained: it fires on old permits or gets lost during a network failure. An administrator who knows the three rules below shortens diagnosis from hours to minutes. Here is the checklist with comments.

13 min
maximum margin of the trigger window beyond the set value
Source: karta
8 min
buffer check cycle
Source: karta

How a time trigger works

A time trigger of a task works in a limited window. It requires at least the set number of minutes from the start of the phase, but no more than 13 minutes beyond it. The upper limit is not an error: it protects against firing tasks backwards on old permits that finished that phase long ago.

For the administrator the rule is simple: the trigger is not an absolute clock, but a window relative to the start of the phase. The configuration therefore depends on both the value of N and the frequency with which the module checks the tasks.

Start of the countdown and the alarm permit

The alarm permit for missing training is generated by the system and does not go through preparation phases. Its creation is the start of the countdown. When configuring tasks for such a permit, do not look for a preparation phase, because there is none.

A common mistake is to define a task relative to a phase that the alarm permit never reaches. Such a task never fires, and the administrator looks for the cause in the network or in the channel.

Time parameters of notifications

Time parameters of notificationsBar chart: trigger window margin 13 minutes, buffer cycle 8 minutes, aggregation window 55 minutes.0204060minTrigger window margin: 13 min13 minTrigger window marginBuffer cycle: 8 min8 minBuffer cycleAggregation window: 55 min55 minAggregation window
Source: InnerWeb feature documentation

Sent marker and network failure

The sent marker of a time-based task is saved only after a successful call. A brief network failure therefore does not lose the notification for good: the task is picked up at the next check, as long as it still falls within the trigger window.

Here the connection with the first rule appears. If the failure lasts longer than the 13 minute margin beyond the set value, the task falls out of the window and will not fire again. The administrator should know how long the network can be unavailable before the notification is lost, and include this in monitoring.

Aggregation and buffers

For tasks that go to the company coordinator, aggregation works: one call, one e-mail and one channel report per company in a rolling 55 minute window. The sending module checks the buffers every 8 minutes. Match expectations to this: a delay of a few minutes in the notification is the result of aggregation, not of a failure.

The grouped call reads up to 5 surnames, and with more people it adds "and others". If a coordinator reports that some people are missing from the call, check the e-mail and the channel report first.

Review rhythm

Once a month review the time tasks: whether N still matches the plant procedure, whether the coordinator numbers are current and whether the log shows no repeated call rejections. After every change of network or notification channels, run a test on a test permit.

Record the result of the review briefly: date, scope, problems found and decisions. Such a record is evidence for an audit that training notifications are maintained, not only switched on.

Checklist

Video

Frequently asked questions

Why did the task not fire on an old permit?

The trigger works in a limited window and does not go beyond 13 minutes past the set number of minutes from the start of the phase. This prevents firing backwards.

Does a network failure lose the notification?

The sent marker is saved only after a successful call, so a brief failure does not lose the notification. A failure longer than the trigger window can cause it to be skipped.

How often are the buffers checked?

The sending module checks the buffers every 8 minutes. Notifications are aggregated in a 55 minute window per company.

Let us verify the time tasks in your installation

We will go through the N values, the trigger window and a test with a network failure together. Let us arrange a conversation with the administrator.

Let's talk