Constraints & Limitations
We believe in being upfront about what SignalBreak can and cannot do. This helps you make informed decisions and set appropriate expectations.
Coverage Limitations
Provider Coverage
SignalBreak monitors 50+ AI providers across Tier 1–4 classifications. However:
| What We Cover | What We Don't (Yet) |
|---|---|
| Major cloud AI providers (OpenAI, Anthropic, Google, AWS, Azure) | Niche or regional providers |
| Established open-source model hosts | Self-hosted private deployments |
| Enterprise AI platforms | Custom fine-tuned model endpoints |
Requesting new providers: Contact support if you need coverage for a provider not currently monitored.
Source Limitations
We aggregate signals from:
- Official status pages
- Changelog feeds
- API documentation changes
- Security advisories
- Pricing announcements
We do not monitor:
- Social media speculation
- Unverified news reports
- Private community forums
- Competitor marketing claims
Geographic Coverage
SignalBreak monitors global provider status. However:
- Regional outages may have delayed detection if not reported on primary status pages
- Some region-specific compliance requirements may not be fully mapped
Data Freshness
| Data Type | Update Frequency |
|---|---|
| Provider health status | Every 5 minutes |
| Signal detection | Near real-time (1–5 minute lag) |
| Governance scores | Recalculated on workflow changes |
| Evidence packs | Point-in-time snapshots |
Note: There is inherent latency between a provider experiencing an issue and reporting it publicly. SignalBreak cannot detect unreported incidents.
Scoring Limitations
Risk Scores
Risk scores are indicative, not definitive. They:
✅ Provide relative prioritisation between events
✅ Account for your specific workflow configuration
✅ Help focus attention on high-impact issues
❌ Cannot predict future provider behaviour
❌ May not capture all business context (e.g., contractual terms, insurance)
❌ Should not replace human judgment for critical decisions
Governance Scores
Governance compliance scores:
✅ Indicate alignment with framework control objectives
✅ Identify gaps requiring attention
✅ Support audit preparation
❌ Do not constitute legal compliance certification
❌ May not reflect all organisational controls outside SignalBreak
❌ Should be validated by qualified compliance professionals
What SignalBreak Is Not
| SignalBreak Is | SignalBreak Is Not |
|---|---|
| A monitoring and intelligence tool | A guaranteed early warning system |
| A governance tracking platform | A legal compliance certification |
| A risk prioritisation aid | A replacement for human judgment |
| An audit evidence generator | An auditor or certification body |
AI-Powered Features
SignalBreak uses AI (Claude, GPT-4, Ollama) for signal enrichment and analysis:
What this means:
- Enrichment quality may vary based on signal clarity
- AI-generated summaries should be verified for critical decisions
- LLM providers may experience their own outages affecting enrichment
- We maintain fallback chains, but degraded mode is possible
Data Retention
| Data Type | Retention Period |
|---|---|
| Signals | 12 months |
| Health logs | 6 months |
| Reports | Indefinite (while account active) |
| Audit logs | 24 months |
Data retention is subject to your plan tier and applicable data protection requirements.
Support Limitations
| Support Type | Availability |
|---|---|
| Live chat | Mon–Fri, 9am–6pm GMT |
| Email response | Within 1 business day |
| Phone support | Enterprise plans only |
| Dedicated CSM | Enterprise plans only |
Requesting Improvements
We're continuously expanding coverage and capabilities. If a limitation affects your use case:
- Contact support — We prioritise based on customer demand
- Submit feature requests — Via live chat or email
- Check our roadmap — Some limitations are already planned for resolution
Last updated: January 2026