Notification Channels & Alert Configuration

Alerts tell you when an integration needs attention, without anyone having to watch a dashboard. Each alert watches one application for one kind of activity and emails a notification channel when its condition is met.

Alerting is set up in two places in Studio:

What	Where	Scope
Notification channels: who receives the email	Settings > Organization > Channels	Shared by every workspace in the organization
Alerts, application digests, and alert history	Settings > Workspace > Alerts	The current workspace

Create the channel first. An alert cannot be saved without one.

Before you begin

• To create or edit channels, you need the Admin or User role. Channels is not shown to other roles.
• The application you want to watch is active in the workspace, and its workflows use the activity the alert type monitors. See Choose an alert type.

Create a notification channel

A notification channel is a named list of email addresses. Alerts refer to a channel by name, so when you change who is in the channel, every alert that uses it follows.

1. Select Settings > Organization > Channels.
2. Select Create Channel. The Create Notification Channel drawer opens.
3. Enter a Channel Name. Names must be unique.
4. In Channel Type, select Email. Email is the only channel type.
5. In Email Recipients, enter an address and press Enter. Tab, a comma, a semicolon, or a space also adds it. Pasting a list of addresses adds them all.
6. Select Save.

The Channels table lists each channel's Name, Type, and Recipients. To change one, select its name or select Edit from the row's More options menu. To remove one, select Delete from the same menu and confirm in the Delete Notification Channel dialog.

Choose an alert type

The type is chosen in Activity Type, and it decides what the alert watches and which applications you can pick. The Application Name list shows only active applications with at least one workflow that uses the matching activity.

Type	Watches	Applications offered	Condition
Scheduler	How many times the schedule ran	A workflow triggered by the Schedule event source	A threshold on the run count. See What a Scheduler alert measures.
Batch	Whether every record in a batch was processed	A workflow that uses InitBatch	None to set. Checked automatically about 3 hours after the batch starts.
HTTP	HTTP error responses from outbound API calls	A workflow that uses RESTClient, or a custom activity built from it	Chosen error codes within an active window
SFTP	SFTP file transfer failures	A workflow that uses an SFTP file activity	Failures within an active window
ABS	Azure Blob Storage file transfer failures	A workflow that uses an ABS file activity	Failures within an active window
Messaging	Messages waiting on a queue or topic	A workflow that receives messages with ReceiveMessage	A message waits longer than the time you set

Scheduler alerts have three steps. The other types have two, because they have no repeat-notification settings.

Create an alert

1. Select Settings > Workspace > Alerts. The Alert Configuration tab opens.
2. Select Create Alert. The Alert Configuration drawer opens.
3. Complete Basic Info.
4. Complete Set Alert Conditions for the type you chose.
5. For a Scheduler alert, complete Notification Settings, or leave it off.
6. Select Save.

Basic info

Field	Applies to	What to enter
Alert Name	All types	A name unique in the workspace.
Activity Type	All types	Scheduler, Batch, HTTP, SFTP, ABS, or Messaging. Changing it clears an application that no longer qualifies.
Application Name	All types	The application to watch.
Aggregation Function	Scheduler	Select count(). See What a Scheduler alert measures.
Span Scope	Scheduler	Select any option, for example Root Spans. Required, but not applied to the alert today.

What a Scheduler alert measures

Each time the schedule fires, the application records a run. A Scheduler alert counts those runs inside the evaluation window and compares the count with your threshold. So the question a Scheduler alert answers is "did my schedule run, and how often?"

Use count(). The other functions, min(), max(), sum(), and avg(), need a value to measure, and the alert is not told which one, so they do not give a meaningful result. Span Scope is not applied, so any value gives the same result.

Alert when a schedule stops running

To be told when a schedule misses its runs:

1. In Aggregation Function, select count(). In Span Scope, select Root Spans.
2. In the threshold sentence, select BELOW, enter 1 as the Threshold, and keep IN TOTAL.
3. Set the evaluation window to Rolling, a little longer than the time between runs. For an hourly schedule, select Last 2 hours, so one late run does not raise an alert.
4. Select a severity and your channels.
5. Expand ADVANCED OPTIONS and turn on Alert when data stops coming. Enter the same number of minutes as the window, 120 for 2 hours. This catches a schedule that records nothing at all.
6. Leave the other settings as they are, and select Save.

The alert fires when the schedule has not run in the window. It does not fire when a run happens but fails, because a failed run still counts as a run.

Conditions for Scheduler alerts

The Threshold condition reads as a sentence. The drawer opens with these values selected:

Send a notification when EQUAL TO the threshold(s) IN TOTAL during the last 10 minutes, critical on value = threshold, send to channels

Set each part:

Part	Options
Comparison	ABOVE, BELOW, EQUAL TO, or NOT EQUAL TO
How the threshold must be met	AT LEAST ONCE, ALL THE TIME, ON AVERAGE, IN TOTAL, or LAST
Evaluation window	See the next table.
Severity	critical, warning, or info
Threshold	The number to compare against. Required. The unit is fixed as Unit.
Channels	One or more notification channels. Required.

The default comparison is EQUAL TO, so an alert saved without changing it fires only on an exact match. For "fewer than" or "more than", select BELOW or ABOVE.

The evaluation window has two kinds:

Evaluation Window	Timeframe options	How it counts
Rolling	Last 5 minutes, Last 10 minutes, Last 15 minutes, Last 30 minutes, Last 1 hour, Last 2 hours, Last 4 hours, or Custom with a value and a unit of Seconds, Minutes, Hours, or Days	A fixed-length window that moves forward continuously
Cumulative	Current hour, Current day, or Current month	Grows from a fixed starting point: the Starting At Minute of each hour, the Starting At time of each day, or the Starting On Day of each month, in the selected timezone

Expand ADVANCED OPTIONS for three more settings:

Setting	What it does
How often to check	How often the alert evaluates its data. The form starts at 30 minutes; the unit can be seconds, minutes, or hours.
Alert when data stops coming	When on, notifies if no data arrives for the number of minutes you enter. Off by default.
Minimum data required	When on, the alert fires only when at least the number of Datapoints you enter are present. Off by default.

Conditions for Batch alerts

A Batch alert has no condition to set. Set Alert Conditions asks only for Notification Channels.

When a batch starts, Koodisi records how many records it declared. About 3 hours later, it counts the records that were processed. If any are missing, the channel receives an email that lists the missing record IDs.

Conditions for Messaging alerts

Field	What to enter
Entity Type	Queue or Topic.
Entity Name	The name of the queue or topic. Required.
Subscription Name	Topic only. The subscription to watch. Required.
Environment	Development or Production.
Max idle time (minutes)	How long a message can wait before the alert fires. Defaults to 5.
Notification Channels	One or more notification channels. Required.

Conditions for HTTP, SFTP, and ABS alerts

These alerts watch for failures during an active window instead of a metric threshold.

Field	What to enter
Active Window Start and Active Window End	The daily window in which failures are watched, as HH:mm. Defaults to 00:00 to 23:59.
Active Days	The days the alert is active. All seven are selected by default; Select All and Clear All are in the list. At least one is required.
Timezone	Asia/Kolkata or UTC. Defaults to Asia/Kolkata.
Enabled	On by default.
Error Codes	HTTP only. One or more of 400, 401, 403, 404, 405, 409, 415, 422, 500, 502, 503, 504. At least one is required.
Notification Channels	One or more notification channels. Required.

Notification settings

For Scheduler alerts, Repeat notifications sends periodic notifications while the alert condition remains active. It is off by default. When you turn it on, set how often to repeat and in which states: Firing, No Data, or both.

Manage alerts

The Alert Configuration tab lists alerts with Alert Name, Activity Type, Channels, Status, History, and Action columns.

• The Inactive / Active switch above the table chooses which alerts are listed. Search alerts filters by name.
• To edit, select the alert name, or select Edit from the More options menu. The Edit Alert Configuration drawer opens.
• To pause an alert without deleting it, select Deactivate from the menu. Select Activate to resume it.
• To delete, select Delete and confirm in the Delete Alert dialog. Deletion cannot be undone.

Review alert history

Select the History icon on an alert's row to see that alert's past runs.

The summary shows Failures, Warnings, Skipped, and Total Runs. The table lists Date & Time, Event, Description, Status, Notification, and Acknowledged. By default it shows All Runs from the Last 30 days. Filter it to Errors, Warnings, or Skipped, or to the Last 7 days or Last 90 days.

• Select Acknowledge on a row to record that someone has seen it.
• Select Download to save the current rows as a CSV file.

Select Alert Configuration in the breadcrumb to return to the list.

Set up an application digest

An application digest is a scheduled summary of an application's integration metrics, emailed to one or more channels. It is configured separately for each application.

1. Select Settings > Workspace > Alerts > Application Digest. Every active application is listed, with Not configured in Status until it has a digest.
2. Select Set up on the application's row. The Configure Digest drawer opens.
3. Turn on Enable Digest.
4. Select a Report Frequency, from 1 day to 30 days.
5. Select one or more Destination Channels.
6. Select Save Digest.

To change or remove a digest, use Edit or Delete in the row's More options menu. Turning off Enable Digest keeps the settings but stops the digest; the row then shows Inactive.

How you know it worked

• A new channel appears in the Channels table, and Studio reports Notification channel created successfully.
• A new alert appears in the Alert Configuration table with ACTIVE status, and Studio reports Alert configured successfully.
• A new digest shows Active in the Application Digest table, and Studio reports Application digest configured successfully.

To prove the whole chain works, make the condition happen in a test environment, then confirm the email reaches the channel and the run appears in the alert's history.

When it does not work

Symptom	Cause	What to do
Channels is missing from Settings > Organization	Your role cannot manage channels	Ask an administrator to create the channel
No channels are offered in Notification Channels or Destination Channels	No notification channel exists yet	Create a notification channel, then reopen the drawer
The application you want is not in Application Name	The application is inactive, or none of its workflows uses the activity for the selected Activity Type	Check the alert type against Choose an alert type
Alert name already exists or Channel name already exists	Names must be unique	Choose another name
Please add at least one email address	Email Recipients is empty	Add an address and press Enter
Please select at least one error code or Please select at least one active day	A required list on an HTTP, SFTP, or ABS alert is empty	Select at least one
Failed to create alert, or a message from the server	Studio could not save the alert	Check the fields and try again. If it persists, contact your administrator.
The alert saves but never fires	The comparison or threshold does not match what you meant; EQUAL TO is the default	Edit the alert, check the comparison and threshold, and look at its history
A Scheduler alert does not fire when the schedule stops	The alert uses a function other than count(), or the comparison is not BELOW	Set it up as in Alert when a schedule stops running

Next steps

• Configuring Alerts and Notifications shows how to create the same channels and alerts by asking NxAgent.
• Investigating Integration Health helps you find what went wrong after an alert fires.