View alerts by signal or noise
Alert signal or noise — powered by Rovo — uses AI to classify incoming alerts based on historical patterns, team behavior, and learned noise signals. This helps on-call teams reduce noise and focus on alerts that are more likely to be meaningful.
Because this capability relies on machine learning models, classifications may not always be perfect. The feature is designed to guide your decision-making, not replace it. We encourage teams to review classifications, use their judgement, and adjust labels when needed. Your feedback continuously improves the model over time.
By classifying alerts over time, teams can reduce alert fatigue and focus on the most critical incidents.
Alert classification types
This filtering automatically analyzes incoming alerts and classifies them into three types:
Signal: Alerts that are likely important and require attention.
Noise: Alerts that are likely low-value or repetitive.
Likely noise: Alerts that look similar to known noise patterns but may occasionally matter.
AI-generated classifications
Rovo first provides a predicted classification for each alert. Once you review and confirm the suggestion, it becomes an accepted classification and the corresponding tag. For example, if you confirm, an AI-predicted signal becomes a signal, and an AI-predicted likely noise becomes a noise.
Rovo uses your feedback to improve its predictions over time
This learning is specific to your system and is not shared with other customers. Predictions and improvements to these predictions apply only within your environment.
How does Rovo credit usage work for Alert signal and noise
Alert signal and noise classification uses Rovo credits when you take certain actions. Each billable action includes a per user, monthly Rovo credit allowance of 100 credits.
What counts as a billable action?
Accept signal and noise classification
What doesn’t count as a billable action?
These actions are free to use:
Update classification
View signal and noise classification
Rovo credit allowances will be applied starting December 3, 2026. Organization admins can view and manage their usage in Atlassian Administration. More about Rovo credit usage
When using our services, customers must comply with the Atlassian Acceptable Use Policy.
What to expect if you go over your credit allowance
Organization and billing admins set Rovo credit usage limits and have controls to adjust limits as needed. When you reach your usage limit with extra usage enabled, billable actions continue without interruption. Admins are notified at 80% and 100% usage, and extra usage is charged on your next invoice. More about extra usage
If extra usage is disabled, you won’t be able to view signal and noise classification column in the alert list. Free actions will continue to work.
If you need to use more than your total allowance, you can also consider purchasing usage packs.
Organization admins can track Rovo credit usage in Atlassian Administration.
To view your usage:
Go to Atlassian Administration. Select your organization if you have more than one.
Select Insights, then Platform usage.
From here, you can view your total credit allowance, current usage, app-level trends, user-level usage, and which features are consuming credits.
How alerts are classified
Alert signal and noise evaluate alerts by using multiple sources of information. Each source contributes to understanding whether an alert is an important signal or a low-value noise.
Conceptually, the system looks at a few elements:
Learned patterns from past alerts
The system looks at historical behavior across similar alerts to understand whether an alert tends to be important or benign. It groups alerts with similar summaries, tags, sources, or structures. Read how alerts are grouped.
It observes how often similar alerts required action in the past. For example, were they acknowledged or assigned to someone? Were similar alerts led to incidents, or were they routinely ignored?
The system uses these historical outcomes to estimate whether the new alert is likely to matter.
This helps teams avoid being paged repeatedly for alerts they consistently treat as noise. The system uses this information to classify the alert and may display a short explanation to help you understand why.
Manually update the alert’s classification
You can update classifications on individual or group alerts manually. Updating classifications helps the system learn your team’s patterns and improves future predictions.
To manually update an alert’s classification:
Select the designated classification of the alert or alert group.
Review the reasoning. You can update the classification based on your decision.
Notification behavior
Alert signal and noise don’t change your existing notification settings. Rovo simply adds the appropriate labels (such as signal or likely noise) to alerts, while your configured routing, escalation, and other alert notification rules continue to work exactly as before.
View and manage alert classification in Slack
You can view and update an alert’s classification directly from Slack. This lets responders review context and take action without switching tools.
When a classification already exists
If an alert already has a classification, it appears in the Slack message with a Classification label.
To review or update it:
In the alert message, select See explanation.
Review the explanation, including:
Why the alert was classified this way
Historical patterns and frequency
Suggested action
Choose one of the following:
Accept classification to confirm the suggested classification
Update as noise to override the suggestion
When no classification exists
If the alert doesn’t have a classification yet, you can add one from Slack.
To add a classification:
In the alert message, select Add classification.
Choose one of the following:
Add as Signal
Add as Noise
The selected classification is applied to the alert and used by Rovo to learn your team’s alert patterns over time.
Was this helpful?