1. Product Overview & Introduction

What is Alerting

Alerting is a real-time monitoring system for payment performance metrics.
It helps you detect anomalies or performance issues as soon as they occur, giving you visibility and control over your most critical payment KPIs.

It provides:

Slack and email notifications when a performance issue is detected
A dedicated Dashboard page to explore and understand alerts in context
Full self-serve capabilities to create and manage your own monitors
A backtesting mode to validate monitor configurations against historical data before going live

Key Use Cases

Alerting helps you monitor key indicators of payment health:

Latency: Track Payment Provider HTTP latency (95th percentile) to detect downtimes
Error Rate: Track Payment Provider related internal errors to detect technical issues (gateway.internal-error, gateway.unknown-error, gateway.timeout, gateway.upstream-error)
Conversion Rate: Detect performance drops to protect your revenue
Transaction Count: Track payment volumes to detect unusual spikes or drops in traffic

Monitors can be tailored to two distinct classes of issues: fast, high-impact incidents—such as a payment gateway suddenly going down—where short evaluation windows on high-volume data allow instant detection; and silent, long-running anomalies, where longer evaluation periods on lower-volume signals help surface subtle issues that might otherwise remain unnoticed.

2. Monitor Types & Capabilities

Alerting supports two complementary monitor types:

Static threshold monitors which trigger when a metric exceeds or falls below a defined threshold.
Smart monitors powered by a seasonal algorithm that automatically defines an adaptative threshold and evaluation period.

Static Threshold Monitors

Each monitor is defined by a set of configuration parameters:

Parameter	Description
Metric	The metric being tracked (`Conversion rate`, `Payment provider latency`, `Payment provider error rate` or `Transaction count`)
Calculation	How the metric is computed against payment attempts: 'Net', 'Raw', or 'Attempt #' (same semantics as in Analytics)
Condition	The comparison operator used to trigger the alert (`>`, `<`)
Threshold	Numeric value that triggers an alert
Evaluation period	Time interval between two evaluations (default: every 5 minutes)
Recovery periods	The number of periods for an alert to be recovered (default: 1)
Filters	Optional filters to focus on specific data (gateway, currency, transaction metadata, etc.)
Minimum volume	Minimum volume required to trigger an alert (default: 100)

Smart Monitors

Smart monitors are powered by a seasonal algorithm that learns from the last 30 days of data and automatically adapts the alert threshold to the metric's expected behaviour at any given time of day. It also defines an optimal evaluation period to have enough volume per period.

This means you no longer need to define a threshold or an evaluation period. The result: easier configuration and more accurate detection, especially for metrics with strong daily patterns.

Smart monitors are particularly well-suited to Conversion rate and Transaction count, where the value naturally fluctuates over time and a static threshold is hard to set.

Parameter	Description
Metric	The metric being tracked (`Conversion rate`, `Payment provider error rate` or `Transaction count`)
Calculation	How the metric is computed against payment attempts: 'Net', 'Raw', or 'Attempt #' (same semantics as in Analytics)
Condition	The comparison operator used to trigger the alert (`>`, `<`)
Sensitivity reduction	A value between 0 and 1 that loosens the auto-tuned threshold to reduce noise. 0 keeps the algorithm at its default sensitivity; higher values widen the tolerated range and produce fewer alerts.
Recovery periods	The number of periods for an alert to be recovered (default: 1)
Filters	Optional filters to focus on specific data (gateway, currency, transaction metadata, etc.)
Minimum volume	Minimum volume required to trigger an alert (default: 100)

How to choose ?

Use a Static threshold monitor when you have an explicit SLO or business-defined limit (e.g. "latency must stay below 10s").

Use a Smart monitor when the metric naturally varies over time and you want to be alerted on deviations from its normal pattern (e.g., conversion rate or transaction count drops that wouldn't be obvious against a fixed threshold).

Backtesting Mode

Before activating a monitor, you can backtest it against historical data to see how it would have behaved in the past. Adjust thresholds, filters, calculation type, or sensitivity reduction, and immediately see which alerts would have triggered — all without sending a single notification to your team.

This is the recommended way to dial in the sensitivity reduction for Smart monitors and to validate static thresholds. No more tuning in production !

3. Notifications & Actions

Notification Channels

Each monitor can fan out alerts and recoveries to multiple destinations: up to 5 Slack channels and 5 email recipients per monitor. Every notification includes a direct link to the corresponding alert on the Alerting dashboard.

Warning: Slack channels must be on the ProcessOut workspace, contact us if you want to create one !

Deep Dive & Contextual Exploration

From any alert from the dashboard, you can access detailed data views:

Jump to Analytics to analyse trends and segment impact
Jump to Transactions to investigate affected payments in detail

4. Getting Started

Access the Alerting Tab
Go to Alerting tab in your ProcessOut Dashboard to view existing monitors.
Create Your First Monitor
- Click Create Monitor
- Choose between a Static threshold or Smart monitor
- Select a metric and a calculation type
- Configure threshold, filters and volume settings
- Add up to 5 Slack channels and 5 email recipients to be notified
- Backtest the monitor against historical data and tweak the configuration until you're happy with the results
Investigate Alerts
- Use the Alerting dashboard to see open and past alerts
- Click an alert to access Analytics or Transactions for detailed investigation

Disclaimer

Alerting is the first version of our fully merchant-configurable alerting system.
We are actively improving it to make monitoring more accurate, flexible, and actionable.

Your feedback is essential — please reach out through your ProcessOut contact or via the support portal to share your experience or suggestions.