Metrics and Alerts
New Relic APM + NRQL alerts, Sentry error monitoring — configuration and alert types
Monitoring Stack
The SSM.ro platform uses two complementary monitoring and alerting systems:
| System | Role | Integration | Model |
|---|---|---|---|
| New Relic | Live APM, application logs, NRQL alerts on metrics | New Relic agent on dynos; HTTPS, key per account | Shared |
| Sentry | Runtime exception monitoring, email alerting | Sentry SDK in the application; HTTPS, DSN per project | Both |
Enterprise (Dedicated) Model
In the Enterprise (Dedicated) Model, application logs are stored in AWS CloudWatch Logs instead of New Relic. Sentry remains active in both models for runtime exception monitoring.
New Relic
What is monitored
- Live application logs — all logs from web dynos and worker dynos
- APM (Application Performance Monitoring) — response time, throughput, error rate
- Ping Monitor — availability check for the main endpoint (
ssm.ro) with alerting on unavailability
NRQL Alert Types
Alerts are configured through NRQL alert policies (New Relic Query Language) on:
- Application error rate — anomaly detection signals at the application level
- Ping Monitor — Critical priority alert when the query returns > 0.0 (endpoint unavailable)
New Relic Alert Example
ssm.ro query result is > 0.0 on 'Ping Monitor'
→ Critical priority issue is active
→ Notification: Acknowledge / Close / Go to issueSentry
What is monitored
- Runtime exceptions from the web application (production and any other configured environment)
- Each captured exception includes: error type, message, stack trace, environment, timestamp, unique ID
Sentry Alert Types
Alerts are sent by email when a new issue appears or on regressions:
| Alert type | Example |
|---|---|
| RuntimeError | RuntimeError: Invalid CLM role (Personal::Aplicatii::TeamController#show) |
| TypeError | TypeError: Permissions check failed |
Alerts include: Project, Environment, Level (error/warning), exception details, and a direct link to the issue in Sentry.
Sentry Alerting Flow
Exception in production
│
▼
Sentry captures the event
│ automatic email
▼
The team receives an alert with full details
│
▼
Triage and remediation per patch management SLALog Coverage
| Source | Destination | Retention | Model |
|---|---|---|---|
| Live application | New Relic | Per New Relic plan | Shared |
| Live application | AWS CloudWatch Logs | Per configuration | Dedicated |
| Runtime exceptions | Sentry | Per Sentry plan | Both |
| Document signing audit | AWS CloudWatch Logs | 5 years | Both |
| Customer (tenant) activity | Platform database | 3 months (downloadable by tenant) | Both |
Intrusion Detection Signals
New Relic NRQL alerts on application error rate provide application-level intrusion detection signals — a sudden spike in errors may indicate attack attempts or unexpected activity. This is one of the compensating controls in place of a dedicated WAF.