Beyond observability: Using reliability scores to drive results
summary
Argues that teams need a measurable reliability score applied across services, not just retrospective alerts, and describes a three-step strategy: establish a baseline, remediate and retest, then automate reliability scores in the pipeline. Shows Gremlin’s reliability management product that runs standardized probes for host and zonal failures, memory exhaustion and network partitions, surfaces per-service scores in a single pane, and supports regular cadence testing (daily/weekly) so improvements are verifiable before incidents occur.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 0:42Everyone cares about reliability
- 1:20Three key challenges
- 2:39Going fast
- 4:35Something greater
- 6:09Three strategic components
- 8:32I need a strategy
- 8:57Gremlins Reliability Management
- 10:36Gremlins Reliability Scores
- 12:23Outro
related talks
26mReliability Scoring: A 3-Part Formula for Promoting Reliable Code
22mReliability Reviews in the Wild: Using Data to Drive Production Health
27mContinuous Reliability. How?
37mBetter Reliability Through Observability and Experimentation
18mProduct Reliability in
35mSRE 101
36mThe road to world-class observability
18mMeasuring Reliability Culture to Optimize Tradeoffs: Perspectives from an...