Alerts

ServersPulse monitors your server metrics in real time and creates incidents when conditions you define are met. You configure these conditions through notification rules on each server's Rules tab.

How alerts work

The plugin sends metric snapshots at your configured interval
The alert engine evaluates each snapshot against your active rules
When a rule's conditions are met for the required duration, an incident is created
Notifications are sent to your configured channels (Discord, webhooks)
When the condition clears, the incident is automatically resolved

Notification rules

Each server can have multiple notification rules. Rules are created and managed from the Rules tab on your server's dashboard page.

Rule types

Metric threshold rules

Define custom conditions based on server metrics. You choose the metric, operator, threshold, and how long the condition must persist before triggering.

Available metrics:

Metric	Key	Description
TPS	`tps`	Ticks per second (target: 20.0)
MSPT	`mspt`	Milliseconds per tick
Memory %	`memory_pct`	Heap memory utilization percentage
Online Players	`online_players`	Current player count
Entity Count	`entity_count`	Total loaded entities

Available operators:

Operator	Meaning
`lt`	Less than
`gt`	Greater than
`eq`	Equal to
`neq`	Not equal to
`lte`	Less than or equal to
`gte`	Greater than or equal to

Example: Alert when TPS is below 18 for 60 consecutive seconds.

Preset template rules

Pre-built rule templates for common scenarios. These use specialized evaluation logic that goes beyond simple metric thresholds:

Template	Description	Default threshold
`player_spike`	Sudden increase in concurrent players	40 players
`zero_players`	Server has zero online players	0 players
`disconnect_burst`	Multiple players disconnect in a short window	10 disconnects
`recovery`	Server TPS recovers above threshold after a drop	TPS > 19

Preset thresholds can be customized when creating the rule.

Rule configuration options

Setting	Description
Name	A descriptive label for the rule
Duration	How many seconds the condition must persist before triggering (`durationSeconds`)
Window	The evaluation time window in seconds (`windowSeconds`)
Cooldown	Minimum seconds between repeated alerts for the same rule (default: 300 seconds / 5 minutes)
Severity	`warning` or `critical` -- affects how the incident is displayed and routed
Enabled	Toggle the rule on or off without deleting it

Condition builder

For metric threshold rules, you can build complex conditions using the condition tree:

Single condition -- One metric/operator/threshold check
Condition groups -- Combine multiple conditions with all (AND) or any (OR) logic
Nested groups -- Groups can contain other groups for complex logic

Example: Alert when TPS < 18 AND memory > 85%, or when MSPT > 50.

Spark auto-profiling

When enabled on a rule (sparkProfile: true), ServersPulse automatically requests a Spark performance profile when the rule triggers. This captures a detailed profiling snapshot at the moment of the incident, giving you immediate diagnostic data.

Spark profiling requires:

The Spark plugin/mod installed on your Minecraft server
A Starter plan or above

The resulting Spark profile URL is attached to the incident for easy access.

Webhook notifications

Rules can optionally include a webhook URL. When the rule triggers, ServersPulse sends an HTTP POST request to the URL with the incident details as JSON. This lets you integrate with external tools, custom dashboards, or automation systems.

Discord channel override

Individual rules can target a specific Discord channel by setting a channel ID, overriding the guild's default alert channel. This lets you route critical alerts to a dedicated channel while sending warnings elsewhere.

Incident types

When rules trigger (or the built-in detection engine fires), incidents are created with one of these types:

Type	Description
`tps_drop`	TPS fell below the configured threshold
`crash`	Server stopped sending data unexpectedly
`memory_high`	Memory utilization exceeded the threshold
`offline`	Server is not reachable or not sending snapshots
`player_spike`	Sudden increase in player count
`zero_players`	Server has no online players
`disconnect_burst`	Multiple players disconnected in a short window
`recovery`	Server performance recovered after an incident
`join_failure`	Players failed to join the server

Severity levels

Each incident has a severity that affects display priority and notification urgency:

Warning -- Performance degradation or notable events that may not require immediate action
Critical -- Severe issues that likely need immediate attention (crashes, offline, extreme lag)

Viewing incidents

Incidents are listed on the Incidents tab of each server page. Each incident shows:

The incident type and severity
The rule that triggered it (if rule-based)
When it started and when it resolved (if resolved)
A descriptive message with relevant metric values
A link to the Spark profile (if Spark profiling was triggered)

Recommended starter rules

If you are setting up alerts for the first time, these are good starting points:

Rule	Metric	Operator	Threshold	Duration	Severity
Low TPS	`tps`	`lt`	18	60s	critical
High MSPT	`mspt`	`gt`	50	60s	warning
Server offline	--	--	--	--	critical
High memory	`memory_pct`	`gt`	90	120s	warning
Player spike	preset	--	40	--	warning

Reducing alert noise

Use sustained durations -- Require conditions to persist for 30-120 seconds to filter out momentary spikes
Set appropriate cooldowns -- The default 5-minute cooldown prevents repeated alerts for the same ongoing issue
Split severity levels -- Use warning for early indicators and critical for things that need immediate response
Disable during maintenance -- Toggle rules off or use the Discord bot's silence command before planned restarts
Review after incidents -- After each major incident, check whether your rules triggered appropriately and adjust thresholds

Contributors

Thanks to everyone who keeps this docs area accurate.

bwmp