An enterprise digital healthcare platform on Google Cloud ingests real-time patient telemetry streams via Cloud Pub/Sub and processes them using a microservices pipeline backed by Cloud Spanner. The contract specifies a customer SLA of 99.9% availability, defined as telemetry processing latency staying under 1.5 seconds. During peak events, while the frontend HTTP Load Balancer immediately acknowledges incoming requests with HTTP 200, downstream processing queue backpressure causes ingestion processing times to frequently exceed 3 seconds. Consequently, the company incurs contract financial penalties despite internal metrics showing 99.99% HTTP request success rates. Which TWO actions should the Site Reliability Engineering (SRE) and business teams take to align technical service levels with business objectives and prevent SLA breaches?
- Define the Service Level Indicator (SLI) as the proportion of valid telemetry events whose end-to-end processing latency from ingestion to database write is under 1.5 seconds, and set the internal Service Level Objective (SLO) at 99.95%.Cevap
- BDefine the Service Level Indicator (SLI) as a targeted goal of 99.9%, and configure the Service Level Objective (SLO) as the real-time proportion of HTTP 200 OK responses returned by the frontend load balancer.
- Establish an error budget policy that automatically halts new feature deployments and shifts engineering focus to database throughput optimization whenever the multi-window burn rate indicates the error budget will deplete before the end of the compliance cycle.Cevap
- DConfigure static operational alerts that immediately page on-call engineers for every single telemetry event that exceeds 1.5 seconds processing time to trigger manual regional database failovers.