SRE for [cyber]security
summary
Describes running low-level and customer-facing security at hyperscale, claiming mitigation of a DDoS at 47 million requests per second and another event near 400 million packets per second, and shows concrete practices: eBPF-based observability, consolidated asset inventory, automated playbooks to prevent human error, and SLO-style security metrics. Emphasizes backups that actually restore, tradeoffs in key management, the need to exercise a cold start emergency mode, and a pragmatic build-vs-buy approach to reduce tooling fragmentation.
AI-written summary · about
chapters (12)
- 0:00SRE in cybersecurity intro
- 3:43Security teams' roles
- 6:32Security observability foundations
- 9:21Backup and disaster recovery
- 12:09SLOs for security
- 16:02SRE evolution to resilience
- 21:18Researching rare security events
- 24:46Security vs availability in keys
- 26:32Disaster resilience modes
- 30:22Learning from hyperscalers
- 33:07Org design for DevOps SRE security
- 34:10Key takeaways
related talks
39mWhen SRE and Security Teams Meet to Face a Crisis
33mSecurity and SRE: Natural Force Multipliers
28mDefending against SRE with SRE
32mIs the S in SRE for “Security”?
35mRole of SRE in DevSecOps
30mIntegrating cloud native security into the SRE culture
20mImplementing SRE in a Regulated Environment
40mCyber Resilience: Say What?