Cloudflare SLA: What the Uptime Guarantee Covers
Cloudflare offers SLA-backed service credits on Business and Enterprise plans, but claims must be filed within five business days.
Security & Outages
Security incidents, outages, and enterprise risk in the software stack.
Key companies & products
Guides & explainers
AWS SLA Guide: What Uptime Credits Actually Cover
AWS offers SLAs for paid, generally available services, but Amazon EC2 credits apply to future bills rather than cash refunds.
Latest news & analysis
AI Agents Credential Theft: Google’s Six-Hour Warning
Google threat reporting says attackers harvested thousands of credentials in under six hours after a cloud compromise, tightening response windows.
Cloudflare Bot Management Adds Adaptive Intelligence
Cloudflare launched Adaptive Intelligence on August 31, 2026, starting with a continuously retrained detection model inside Cloudflare Bot Management.
Veeam Cyber Resilience Ties AI Risk to Recovery
Recent coverage linked Veeam’s AI-risk stance to recovery readiness, and said Data Platform v13.1 and Vault Archive are now generally available.
Cloudflare OS Launch Puts Agent Governance in Focus
Cloudflare OS is now available as an open-source AI workspace, with deployment in an organization's own Cloudflare account available now and a managed dashboard option coming soon.
Agentic Threat Prevention: Cato Predicts AI-Assisted Attack Paths
Cato Networks launched Cato Agentic Threat Prevention on Aug. 3, 2026, a capability that uses autonomous agents to predict attack paths on its SASE platform.
Anthropic Claude security testing exposes enterprise eval risk
Anthropic said it reviewed 141,006 Claude evaluation runs and found three breaches during internal cyber tests, including a PyPI package incident.
Microsoft Agentic Security Platform: Project Perception Launch
Microsoft launched Project Perception and MAI-Cyber-1-Flash, tying agent-based defense to its MDASH vulnerability-management harness and about 96% CyberGym claims.
1Password Claude Integration Adds AI Access Controls
1Password for Claude gives Mac users session-based credential access while keeping passwords and MFA codes out of Claude and Anthropic’s systems.
Prompt Injection Testing for AI Agents: OpenAI’s GPT-Red
OpenAI disclosed details of GPT-Red, an internal system to find prompt-injection flaws at scale, with reported gains in GPT-5.6 Sol defenses.