AI Monitor — Anomaly Report

Period: 2026-09-28 to 2026-10-05 (7 days) • Generated: 2026-10-05 18:37 UTC (2026-10-05 14:37 (America/New_York))

Executive Summary

This week's monitoring detected 732 anomalies with 618 classified as high severity, predominantly driven by a single metric (LogProcessor/processor/PatternDetected at 518 occurrences) that warrants immediate investigation to determine if this represents a genuine system issue or a detection tuning problem. While most metrics show stable or improving trends, the concentration of anomalies in log pattern detection combined with elevated response times on the Application ELB and message queue age suggests potential downstream processing bottlenecks that should be prioritized for root cause analysis. Recommend reviewing LogProcessor configuration and alert thresholds to reduce false positives while investigating whether the SQS and Firehose anomalies indicate capacity

Summary

MetricValue
Total Anomalies Detected732
High Severity (score ≥ 7)618
Medium Severity (score ≥ 5)25
Low Severity (score < 5)89
Suppressed (maintenance)6
Active Subscriptions20
Silent Subscriptions (email disabled)5
Maintenance Windows Configured1

Trends

TrendSubscriptions
Improving2
Stable4
New2

Top Anomalous Metrics

MetricCount
LogProcessor/processor/PatternDetected518
AWS/ApplicationELB/TargetResponseTime50
AWS/Firehose/DeliveryToS3.DataFreshness49
AWS/SQS/ApproximateNumberOfMessagesVisible48
AWS/SQS/ApproximateAgeOfOldestMessage35
AWS/Lambda/Throttles14
AWS/EC2/CPUUtilization12
AWS/ES/IndexingRate5
AWS/ApplicationELB/HTTPCode_ELB_5XX_Count1

Anomalies by Day of Week

DayCount
Mon235
Tue2
Wed24
Thu30
Fri1
Sat101
Sun339

Top Dimensions (Resources)

DimensionCount
PatternCategory=SQL518
PatternName=sql_drop518
PatternSeverity=high518
QueueName=AISIEM-flow-ingestion54
DeliveryStreamName=AIMonitor-metrics49
LoadBalancer=app/AISIEM-siem-alb/fbc6b3ddb022963026
QueueName=AISIEM-threat-detection23
LoadBalancer=app/logprocessor-dashboards-alb/5d8c545e3ef0565018
TargetGroup=targetgroup/AISIEM-editor-tg/459e8e737c17b80514
InstanceId=i-008793ffe280f5fbd8

Anomalies by Account

AccountCount
111111111111732

Cost Analysis (Daily Spend)

ServiceAccountToday7-day AvgChange
AmazonVPC111111111111$1.44$2.51-43%
AmazonOpenSearchService111111111111$1.41$4.01-65%
AmazonCloudWatch111111111111$1.21$2.82-57%
AWSSupportEssential111111111111$0.41$0.00new
AWSELB111111111111$0.36$0.63-43%
AmazonS3111111111111$0.13$0.27-52%
AWSLambda111111111111$0.09$0.20-56%
awskms111111111111$0.05$0.10-48%
AmazonEC2111111111111$0.05$0.08-41%
AWSMarketplace111111111111$0.02$0.01+100%
AmazonDynamoDB111111111111$0.01$0.02-52%
AWSQueueService111111111111$0.01$0.02-48%
AmazonKinesisFirehose111111111111$0.01$0.02-53%

Maintenance Windows

NameTypeDaysStartDurationAffectedSuppressed
WeeklyScheduledsun05:00 UTC / 01:00 (America/New_York)30mall0

Baseline Training Status

Training reflects how much historical data each subscription has collected. Ready subscriptions score anomalies accurately; Learning subscriptions may have elevated false-positive rates early on.

StatusCount
Ready (scoring)25
Learning (not yet scoring)0
No data (metric not streaming)5
Ready (25)
AIMonitor Billing
ALB 5xx errors
ALB p95 response time
ALB target 5xx (backend errors)
Billing EstimatedCharges
EC2 CPUCreditUsage
CloudWatch Firehose Delivery
Monitor successful deliveries to S3 to ensure data is being persisted
Firehose FailedValidation.Bytes
Monitor API call volume from remote account metric stream to detect unusual activity patterns
EC2 CPU utilization (all instances)
LogProcessor/processor PatternDetected (SQL)
LogProcessor/processor PatternFiltered
Firehose delivery lag (delivery backing up)
Lambda concurrency (approaching reserved limit)
Lambda max duration (timeout risk)
Lambda errors (all functions)
Lambda throttles (all functions)
OpenSearch cluster CPU utilization
OpenSearch free storage (running out)
OpenSearch indexing rate (sudden drop = ingestion problem)
OpenSearch JVM memory pressure
SQS message age (processing backlog)
SQS queue depth (backpressure)
SNS failed notification deliveries
No data (5) — metric has not emitted data, verify Metric Stream filter or disable if not needed
API Gateway 5xx errors
API Gateway p95 latency
API Gateway Count
CloudFront 5xx error rate
LogProcessor/processor PatternDetected

AI Monitor System Health

CheckValueStatus
Detector Lambda errors (24h)0OK
Collector Lambda errors (24h)0OK
Firehose Transform errors (24h)0OK
Detector throttles (24h)0OK
Detector invocations (24h)300OK
Billing Processor errors (24h)0OK
Billing Processor invocations (48h)48OK
Discovery Lambda errors (24h)0OK
Discovery Lambda invocations (24h)6OK