SRE -- culture and strategy @ TrustingSocial
summary
Company runs across five data centers and three clouds, ingesting 1 TB metrics/week and scaling an SRE bench from 4 to 15 in nine months. Concrete claims: metrics from air‑gapped systems via reverse hole punching, Postgres ingestion overload solved by pushing data to multiple stores, DNS and Squid proxy failure modes observed, and stateful storage using port works. Built tooling for safe infra changes and access control — Tessellate and Escher — plus extra/LDAP/Nomad workflows and TS2FA OTP gating.
AI-written summary · about
chapters (13)
- 0:00Start
- 0:14SRE: Culture and Strategy
- 5:09Infrastructure Management
- 6:08Managing concurrent access - Retries and rollbacks - Maintain versions
- 6:53Data Model
- 10:19Control the Access and How!
- 10:43Homogeneity: - Making on call manageable - Making deployments faster
- 15:55LDAP for Visibility
- 16:50LDAP for Debuggability
- 17:15Networks and all that jazz.
- 20:34Shared Storage
- 22:38Revisit: Essence of SRE
- 24:26These Tools are Open Source!
