Alert Manager configures and manages all alerts in one place. It leverages the current Smart Alerts framework coupled with simplicity, scalability, and flexibility. One example of a new way to create a Smart Alert using the Alert Manager is to do so for combinations of topics and themes in listening. For detailed information, see – Sprinklr Alert Manager.

A new entry is made in the list view with the details. The status will be set as "in progress" while the setup goes on and changes to "active" once done. After that, the subscriber will get notifications whenever an anomaly occurs.

Setting the "Every time" frequency triggers an email every time there is a spike in the volume of mentions above your preset volume threshold. E.g., If you select Volume of Mentions greater than or equal to 50 and select Alert Frequency as Every time, the system will send an email alert as soon as the volume of mentions crosses 50.

Custom Volumetric Alert is available for the Listening module only.

No. You can only check Custom Volumetric Alerts via email notifications.

• For Social Reporting:

  • Post Insights

  • Account Insights

Refine Alerts is an optional section that allows you to set conditions to get notifications for specific alerts. There are seven conditions to refine alerts, and each module has one or more of these conditions available. The details are as follows:

#1 – Mentions

Subscribe to alert notifications based on mentions. An alert is triggered if the number of mentions exceeds a preset value. Mentions is also available as a filter in the Rule Engine conditions for Automatic Alerts. Mentions condition is not applicable for Engagement Alerts.

#2 – Alert Severity

Subscribe to alert notifications as per the severity defined on a scale of 1 to 10. It only takes numeric values from 1 to 10. Severity measures the variance of the detected alert from the expected value. The higher the variance, the more the severity. Alert Severity is not applicable for Engagement Alerts.

For example, each alert has a severity value associated with it. You create an alert on a topic and add a condition – Alert severity > 5. A Smart Alert is generated every time the system detects an anomaly for this topic and is visible in the engagement column. But you will get an email notification only when the severity of the detected alert is greater than 5. You can also define different segments of users depending on the severity of alerts they should receive.

Find below the various cases of Alert Severity and how it works:

• Alert Severity greater than 6: This will send an email notification only when the severity of the generated alert is greater than 6. By this, you inform the system to send you an email only when the generated alert is too severe.

• Alert Severity less than 6: This will send an email notification only when the severity of the generated alert is less than 6. By this, you inform the system to send you an email when the generated alert is moderately severe.

#3 – Alert Longevity

It is used to measure the persistence of an issue. You can choose to get an alert only if there are more than a specific number of alerts on the same topic or topic group. If longevity is > 2, you will get an email when Sprinklr detects three alerts for the specified condition.

For example, you create an alert on a Topic and add a condition – Alert longevity > 2. You will get an email notification only when more than two anomalies are in the next 24 hours. It helps you ensure that you get emails only when an alert for the same condition occurs more than n times, indicating the persistence of the issue. You can see below the longevity of an alert is two as there are two red spikes.

Find below, the various cases of Alert Longevity and how it works:

• Alert Longevity less than 2: You will get an email notification when the first anomaly is detected. The next email will be sent after 24 hours, i.e., the cooldown period.

• Alert Longevity greater than 2: You will get an email notification only when more than two anomalies are detected for the same condition in the cooldown period.

• Alert Longevity equal to 2: You will get an email notification only when two anomalies are seen for the same condition in the cooldown period.

#4 – Alert Severity Increase

Set Alert Severity Increase as True to get alerts only when the severity level is higher than the previously received alert. Alert Severity Increase is not applicable for Engagement Alerts.

#5 – Total Retweets

Choose to get an email notification only when a particular tweet has more than a specified number of retweets. Total Retweets is applicable for Engagement Alert.

#6 – Retweets Gained

Retweets Gained is the total number of retweets received between the previous and current periods. If there were ten retweets in the last period and forty in the current period, the retweets gained would be 30. You can choose to get a notification only if retweets gained are more, less, or equal to a particular number. Retweets Gained is applicable for Engagement Alert.

#7 – Alert Type

It is relevant only for Location Insights and Product Insights. It can have two values – High or Low, depending on whether you want to get alerts only on spikes or only on drops in critical metrics (star ratings, experience scores, mentions, and insights). If nothing is selected, you will get an alert on both spikes and dips.

You can use the Refine Alerts option in the Alert Manager. The Refine Alerts option has five conditions that you can use to refine/filter alerts: Mentions, Alert Severity, Alert Longevity, Alert Severity Increase, Total Retweets, Retweets Gained, and Alert Type. For detailed information, refer to the previous FAQ — What are the various options under the Refine Alerts section in Alert Manager?

You can also have limited types of Alert categories set up for email notifications, such as only Negative Sentiment and not all of them: Mentions, Positive Sentiment, Negative Sentiment, Language, Country, and Engagement.

You can set up hierarchies or distinct POCs( point of contact ) for alert management through queues and user groups to restrict the number of users getting email notifications for a specific type of alert.

There are four types of Status:

• In Progress: Once you create a new alert, the data setup for that alert takes a few minutes.

• Active: Once the data setup is complete, the alert is active, and subscriber(s) will get notifications whenever an anomaly is detected.

• Failed: When the data setup fails due to server downtime or a backend job failure. It is rare, but please raise a support ticket if it happens.

• Inactive: Disable an alert by clicking on the toggle button in the list view The status changes to Inactive.

Create an engagement column. For detailed information, see – View Anomalies in Engagement Dashboard.

Active status implies that the setup is complete, and you will start getting email notifications as soon as an anomaly is detected. There is no wait time as the Machine Learning model gets trained immediately using current and past data matching the conditions used in the alert. The anomaly detection is entirely dependent on the data available. You can track anomalies via Engagement Columns. For detailed information, see – View Anomalies in Engagement Dashboard.

We show all past anomalies that are detected in the last 24 hours from the time the current anomaly is detected. Here is an example of how the text looks in the engagement column or in an email:

Past anomalies on this tweet:

• with 62 retweets from 07-09-20 18:30 to 07-09-20 18:45

• with 42 retweets from 07-09-20 18:45 to 07-09-20 19:00