Key Monitoring Metrics
Not Started
Loading...
Monitoring Metrics
Essential metrics and KPIs for system monitoring and observability
15
Key Metrics
Application Performance
criticalhistogram
Response Time
Time taken to process and respond to a request
Formula:
Response Time = Request End Time - Request Start TimeThresholds:
Good
< 200ms
Warning
200ms - 1s
Critical
> 1s
Examples:
API endpoint response timeDatabase query execution timePage load time
Monitoring Tools:
PrometheusNew RelicDataDogGrafana
Recommended Alerts:
- P95 response time > 500ms
- Average response time > 200ms
Application Performance
criticalgauge
Throughput (RPS)
Number of requests processed per second
Formula:
Throughput = Total Requests / Time PeriodThresholds:
Good
> 1000 RPS
Warning
500-1000 RPS
Critical
< 500 RPS
Examples:
HTTP requests per secondDatabase transactions per secondMessage queue processing rate
Monitoring Tools:
PrometheusGrafanaCloudWatchNew Relic
Application Performance
criticalgauge
Error Rate
Percentage of requests that result in errors
Formula:
Error Rate = (Error Count / Total Requests) × 100Thresholds:
Good
< 0.1%
Warning
0.1% - 1%
Critical
> 1%
Examples:
4xx client errors5xx server errorsDatabase connection errors
Monitoring Tools:
PrometheusSentryBugsnagCloudWatch
Recommended Alerts:
- Error rate > 1%
- 5xx errors > 0.5%
Infrastructure
highgauge
CPU Utilization
Percentage of CPU capacity being used
Formula:
CPU Utilization = (CPU Used / CPU Total) × 100Thresholds:
Good
< 70%
Warning
70% - 85%
Critical
> 85%
Examples:
Server CPU usageContainer CPU usageDatabase CPU usage
Monitoring Tools:
PrometheusCloudWatchDataDogGrafana
Infrastructure
highgauge
Memory Utilization
Percentage of memory being used
Formula:
Memory Utilization = (Memory Used / Memory Total) × 100Thresholds:
Good
< 80%
Warning
80% - 90%
Critical
> 90%
Examples:
RAM usageJVM heap usageContainer memory usage
Monitoring Tools:
PrometheusCloudWatchDataDogGrafana
Infrastructure
highgauge
Disk Utilization
Percentage of disk space being used
Formula:
Disk Utilization = (Disk Used / Disk Total) × 100Thresholds:
Good
< 80%
Warning
80% - 90%
Critical
> 90%
Examples:
File system usageDatabase storage usageLog file storage
Monitoring Tools:
PrometheusCloudWatchNagiosZabbix
Database
highgauge
Database Connection Pool
Number of active database connections
Formula:
Pool Utilization = (Active Connections / Max Connections) × 100Thresholds:
Good
< 70%
Warning
70% - 85%
Critical
> 85%
Examples:
PostgreSQL connectionsMySQL connectionsMongoDB connections
Monitoring Tools:
PrometheusDataDogNew RelicCloudWatch
Database
criticalhistogram
Database Query Time
Time taken to execute database queries
Formula:
Query Time = Query End Time - Query Start TimeThresholds:
Good
< 100ms
Warning
100ms - 500ms
Critical
> 500ms
Examples:
SELECT query execution timeINSERT/UPDATE query timeComplex JOIN query time
Monitoring Tools:
PrometheusDataDogNew Relicpganalyze
Database
highcounter
Database Deadlocks
Number of database deadlocks per time period
Formula:
Deadlock Rate = Deadlocks / Time PeriodThresholds:
Good
0 per hour
Warning
1-5 per hour
Critical
> 5 per hour
Examples:
Transaction deadlocksLock wait timeoutsConcurrent update conflicts
Monitoring Tools:
PrometheusDataDogCloudWatchDatabase logs
Network
highhistogram
Network Latency
Time for data to travel between network endpoints
Formula:
Latency = Round Trip Time / 2Thresholds:
Good
< 50ms
Warning
50ms - 100ms
Critical
> 100ms
Examples:
Service-to-service latencyDatabase connection latencyCDN latency
Monitoring Tools:
PrometheusGrafanaPingdomCloudWatch
Network
mediumgauge
Bandwidth Utilization
Percentage of network bandwidth being used
Formula:
Bandwidth Utilization = (Used Bandwidth / Total Bandwidth) × 100Thresholds:
Good
< 70%
Warning
70% - 85%
Critical
> 85%
Examples:
Ingress bandwidth usageEgress bandwidth usageInternal network traffic
Monitoring Tools:
PrometheusCloudWatchDataDogSNMP monitoring
Business
highgauge
Active Users
Number of users actively using the system
Formula:
Active Users = Unique Users in Time WindowThresholds:
Good
> 10,000
Warning
5,000 - 10,000
Critical
< 5,000
Examples:
Daily Active Users (DAU)Monthly Active Users (MAU)Concurrent users
Monitoring Tools:
Google AnalyticsMixpanelAmplitudeCustom metrics
Business
criticalgauge
Conversion Rate
Percentage of users who complete desired actions
Formula:
Conversion Rate = (Conversions / Total Visitors) × 100Thresholds:
Good
> 5%
Warning
2% - 5%
Critical
< 2%
Examples:
Sign-up conversion ratePurchase conversion rateEmail subscription rate
Monitoring Tools:
Google AnalyticsMixpanelOptimizelyCustom dashboards
Security
highcounter
Failed Login Attempts
Number of unsuccessful authentication attempts
Formula:
Failed Login Rate = Failed Logins / Time PeriodThresholds:
Good
< 100 per hour
Warning
100 - 500 per hour
Critical
> 500 per hour
Examples:
Brute force attemptsInvalid credentialsAccount lockouts
Monitoring Tools:
PrometheusSplunkELK StackSecurity logs
Recommended Alerts:
- Failed logins > 1000/hour
- Account lockout threshold reached
Security
criticalcounter
Security Events
Number of security-related events detected
Formula:
Security Event Rate = Security Events / Time PeriodThresholds:
Good
0 per hour
Warning
1-5 per hour
Critical
> 5 per hour
Examples:
Intrusion attemptsMalware detectionPrivilege escalation
Monitoring Tools:
SIEM toolsSplunkELK StackSecurity scanners
Monitoring Strategy
6
critical Priority
8
high Priority
1
medium Priority
0
low Priority
Monitoring Best Practice: Start with Critical and High priority metrics to establish baseline monitoring. Implement alerting for all Critical metrics. Use dashboards to visualize trends and correlations across different metric categories.