๐จ Alarm Rules โ Monitoring Conditions โ
๐ Overview โ
JBL provides a flexible and intelligent alert system to detect user drop-offs, abnormal API response patterns, and ISP quality issues.
Alerts are triggered when specific thresholds are reached, helping teams identify system abnormalities, resource bottlenecks, or potential service disruptions early.
๐ Alert Rule Summary โ
| Rule ID / Type | Trigger Frequency | Alert Condition | Monitoring Target |
|---|---|---|---|
[AR01]Domain User Count | Hourly | Compares current user count with the past two weeks: If both previous weeks > 400 users and current value drops below 15%, then: ๐ข Green: Normal ๐ก Yellow: Abnormal for 1โ2 hours ๐ด Red: Abnormal for 3+ consecutive hours | Detect abnormal drops in site domain user activity |
[AR02]ISP User Count | Hourly | Same as AR01, but calculated based on ISP dimensions | Detect abnormal ISP user losses |
[AR03]Response Time | Every 5 minutes | If the average global response time exceeds 800ms for 3 consecutive intervals | Identify potential API logic errors or server overloads |
[AR04]Response Time + Request Count | Every 5 minutes | If response time increases by over 200% and request count drops by over 50% | Detect severe server performance degradation or failed scheduled tasks |
[AR05]Domain + ISP Error Rate | Every 5 minutes | Applicable during peak usage: When domain users > 200 and 12-hour error rate > 2%: ๐ธ Trigger 1: Connection drops by 50%, error rate increases by 200% ๐ธ Trigger 2: Users > 800, connection drops by 25%, error rate increases by 50% | Detect if the domain is blocked or degraded by specific ISPs |
๐ Terminology โ
- User Count: Number of active and continuously connected IPs
- Response Time: Average latency (in milliseconds) for APIs or pages during the measurement interval
- Request Count: Total number of API requests within a specific time window
- Failure Rate:
(Fail / (Fail + Success)) ร 100