Create a New Monitor

This guide will walk you through the steps of creating a monitor in Sumo Logic, from setting up trigger conditions to configuring advanced settings, notifications, and playbooks.

Our AI-driven alerts use machine learning to analyze historical data, establish baselines, detect significant deviations, and filter out irrelevant alerts to reduce alert fatigue and help teams focus on critical issues. These capabilities apply to both logs and metrics, providing a comprehensive monitoring solution. With seasonality detection and customizable anomaly clustering, false positives are minimized, enabling faster issue resolution.

Integrated playbooks automate incident response by gathering diagnostics, notifying teams, triggering recovery actions, and streamlining workflows to improve response times. You can link playbooks to monitors to automate tasks such as restarting services or scaling infrastructure, ensuring swift and efficient anomaly resolution.

Open the New Monitor window

There are several ways to create a new monitor, depending on where you are in Sumo Logic.

From Monitors

Classic UI. In the main Sumo Logic menu, select Manage Data > Monitoring > Monitors.
New UI. In the main Sumo Logic menu, select Alerts > Monitors. You can also click the Go To... menu at the top of the screen and select Monitors.
Click Add > New Monitor, and the New Monitor dialog box will appear.

From Metrics Explorer

Creating a monitor based on the threshold values defined in the Metrics page can save time and effort. By using the pre-filled monitor editor, you can quickly create a monitor with the same threshold values as defined in the Metrics page. This will ensure that the monitor is using the same criteria as the Metrics page, providing consistency in monitoring.

To create a monitor from the Metrics Explorer, follow the steps below:

Open the Metrics Explorer:
- Classic UI. From Sumo Logic home, click Metrics.
- New UI. Click the Go To... menu at the top of the screen and select Metrics Search.
On the Metrics Explorer page:
1. Enter a metrics query.
2. In the Thresholds section, define the critical and warning thresholds for your metrics query.
Click the three-dot kebab icon button at the end of the query field and select Create a Monitor.
The New Monitor will open with prefilled data based on the threshold values you set in the previous steps.
In the Trigger Type section, enable the checkbox that corresponds to the threshold value that you want to use (Critical and/or Warning).
- The threshold values will be the same as defined in the Metrics page for both Critical and Warning thresholds.
- Set all other parameters to default, including the window (15 minutes) and the at all times box.
- Ensure that the Recover value is set to the default, which is the opposite of the Alert value. The Edit Recovery button should be off.
Once all values have been set, click Save to create the monitor.
The same threshold will also be applied to the histogram chart.

note

The same threshold translating functionality supports opening the Alerts Response Page in the Metrics Explorer and opening a monitor in the Metrics Explorer.

tip

When you create a monitor and open the metrics search query in the Metrics Explorer, the signal gets a new value for the quantize operator based on the time range of the query. The default value for the quantize operator is 1m. Because opening the query in Metrics Explorer may not match because of quantization differences, you may need to adjust the query to return the results you expect, especially when creating a monitor that uses the anomaly detection method.
Metrics query with quantize

Step 1. Set trigger conditions

The first step when creating a new monitor is setting the Trigger Conditions.

Monitor Type

Select a Monitor Type, which will create alerts based on Logs, Metrics, or an SLO.
Monitor types

Detection Method

Next, select a Detection Method (not applicable to SLO monitors).

Static

Set specific threshold conditions for well-defined KPIs with constant thresholds (for example, infrastructure metrics like CPU utilization and memory).

Anomaly

Leverage machine learning to identify unusual behavior and suspicious patterns by establishing baselines for normal activity. This AI-driven alerting system uses historical data to minimize false positives and alerts you to deviations.

Model-driven detection. Machine learning models create accurate baselines, eliminating guesswork and noise.
AutoML. The system self-tunes with seasonality detection, minimizing user intervention and adjusting for recurring patterns to reduce false positives.
User-defined sensitivity. Users set alert sensitivity and thresholds, providing context to filter out noise.
One-click playbook assignment. Monitors automatically link to Sumo Logic Automation Service playbooks, expediting incident response.
Auto-diagnosis and recovery. The Automation Service handles diagnosis and resolution, closing the loop from alert to recovery.
Customizable detection. Use advanced rules like "Cluster anomalies" to detect multiple data points exceeding thresholds within a set timeframe.

Use Outlier

If you want to trigger alerts on outlier direction rather than anomaly detection, select Anomaly and enable Use Outlier. This detects unusual changes or spikes in a time series of a key indicator. Use this detection method when you are alerting on KPIs that don't have well-defined constant thresholds for what's good and bad. You want the monitor to automatically detect and alert on unusual changes or spikes on the alerting query. For example, application KPIs like page request, throughput, and latency.
Screenshot of the Monitor Type and Detection Method options in Sumo Logic's 'New Monitor' setup page. Logs is selected as the Monitor Type, and Anomaly is selected as the Detection Method. There is an option to use Outlier detection, which is currently toggled off.

Anomaly or Outlier Direction

If you choose an anomaly or outlier detection method, you'll need to select the Anomaly Direction or Outlier Direction you want to track (not applicable to static detection method).

Up. Get alerted if there is an abnormal increase in the tracked key indicator.
Down. Get alerted if there is an abnormal decrease in the tracked key indicator.
Both. Get alerted if there is any abnormality in the data whether an increase or a decrease.

Query

tip

For guidance on optimizing scan costs when using Flex Pricing, refer to the FAQ on optimizing scan costs for monitors.

In this step, you'll need to provide a logs or metrics query. This is not applicable to SLO monitors.

Logs

Logs monitors can have one query up to 15,000 characters long.

Metrics

Metrics monitors can have up to 6 queries. If you're providing multiple metrics queries, use the letter labels to reference a query row. The monitor will automatically detect the query that triggers your alert, and will mark that row with a notification bell icon. See Join metrics queries for details.
Screenshot of the 'New Monitor' setup page in Sumo Logic, showing the Trigger Conditions section. Metrics is selected as the Monitor Type and Static as the Detection Method. The query includes two metrics: CPU_Sys and CPU_User, with an alert condition combining both metrics (#B + #C). A bell icon is highlighted on the left side.

Trigger Type (Logs)

You can set a logs monitor trigger to alert based on the following:

A returned row count (default), which is the number of rows returned from the log search.
A numeric field returned from the search. You can pick any numeric field from your query, and alert on the value of that field. The field is _count in the above screenshot. To convert a string to a number use the num operator. For example, if you have a field named duration, you would use the num operator as follows to convert it to a number value.
```
| num(duration)
```

Triggers are evaluated by balancing the requirement of timely alert notifications while ensuring that monitor data is indeed available to evaluate trigger conditions.

For static logs monitors, triggers are similar to "Alert when the result is greater than _ within Y Minutes". The triggers are evaluated periodically as below.
When detection window (Y) is Evaluate trigger every
15m or less 1m
15m to 1h 2m
1h to 6h 10m
Greater than 6h 20m

When detection window (Y) is	Evaluate trigger every
15m or less	1m
15m to 1h	2m
1h to 6h	10m
Greater than 6h	20m

For anomaly logs monitors, triggers are evaluated every timeslice as specified in the monitor query. For example, the below query is evaluated every 2 minutes.

_sourceCategory=Labs/Apache/Access
| timeslice 2m
| parse "HTTP/1.1\" * " as status_code
| if (status_code = "200", 1, 0) as successes
| if (status_code = "404", 1, 0) as fails
| sum(successes) as success_cnt, sum(fails) as fail_cnt by _timeslice
| (fail_cnt/(success_cnt+fail_cnt)) * 100 as failure_rate_pct

For outlier logs monitors, triggers are evaluated every 5 minutes.

When configuring monitor trigger conditions, you can set a resolution window to resolve alerts quickly once the underlying issue is fixed. The resolution window specifies how long a monitor will wait before resolving an alert after the issue is corrected.

For example, if your monitor evaluates the last 1 hour, you can set a resolution window of 15 minutes. Once the resolution window is continuously satisfied for 15 minutes, the alert will resolve automatically.
config-resolution-window

Static detection method

Example: Logs - Static - Critical and Warning

Alert when result is <threshold type> <threshold> within <time range>

Parameter	Description
`<threshold type>`	How you want the value compared. Select greater than, greater than or equal, less than or equal, or less than.
`<threshold>`	The value against which the trigger will be evaluated. You can specify any valid numeric value up to 1,000.
`<time range>`	The duration of time to evaluate (values range from 5 minutes to 24 hours).

The recovery condition is set by default to the opposite of the alert condition. If you need to change these settings, switch on the Edit recovery settings toggle and then adjust values for the recovery settings accordingly.

For example, if an alert is set to greater than 10, the recovery would be set to less than or equal to 10 when inferred. Sumo Logic automatically resolves the incident when the resolution condition is satisfied.

Example: Logs - Static - Missing Data

Alert when missing data within <time range>

Parameter	Description
`<time range>`	The duration of time to evaluate (values range from 5 minutes to 24 hours).

For recovery, Sumo Logic will automatically resolve the incident when the resolution condition is satisfied.

Anomaly detection method

Example: Logs - Anomaly - Critical

Alert when anomaly count is at least <count> (max. 1) at any time within <time range>

Parameter Description

<count> Enter the minimum number of anomalies to detect during the detection window before triggering an alert. Ensure that the time period window is 5-10 times longer than the timeslice used in the log query. This setting helps you add context to anomaly detection. For example, if you know a particular signal is noisy, you may want to wait for a number of anomalous data points in the detection window before triggering an alert. If the time period is set to 5 minutes, and the minimum anomaly count is set to 1, then an alert is triggered if 1 anomaly appears within a 5-minute time period.

<time range> The duration of time to watch for anomalies (values range from 5 minutes to 24 hours).

Parameter	Description
`<count>`	Enter the minimum number of anomalies to detect during the detection window before triggering an alert. Ensure that the time period window is 5-10 times longer than the `timeslice` used in the log query. This setting helps you add context to anomaly detection. For example, if you know a particular signal is noisy, you may want to wait for a number of anomalous data points in the detection window before triggering an alert. If the time period is set to 5 minutes, and the minimum anomaly count is set to 1, then an alert is triggered if 1 anomaly appears within a 5-minute time period.
`<time range>`	The duration of time to watch for anomalies (values range from 5 minutes to 24 hours).

Tune the number of anomalous data points detected per day compared to the predicted baseline for the detection window. Select more alerts if you do not want to miss out on most anomalies.

Outlier detection method

Example: Logs - Outlier - Critical and Warning

Alert when result is greater than or equal to <threshold> standard deviations from baseline for <consecutive> consecutive out of <window> data points

Parameter	Description
`<threshold>`	The number of standard deviations for calculating violations. The default is `3`.
`<consecutive>`	The required number of consecutive indicator data points (outliers) to trigger a violation.
`<window>`	The number of data points used to calculate the baseline for outlier detection.

The recovery condition will always be the opposite of the alerting condition. For example, if there is no outlier identified for the duration of the detection window from the time the alert was first fired, then the monitor will be brought back to the normal state. You cannot customize the resolution condition for the monitor.

Example: Logs - Outlier - Missing Data

Alert when missing data within <time range>

Parameter	Description
`<time range>`	The time span of data to evaluate (values range from 5 minutes to 24 hours).

For recovery, Sumo Logic will automatically resolve the incident when the resolution condition is satisfied.