slo-management

# Elastic SLO Management ## Checking SLO Status List all SLOs and their current status: ```bash elastic slos list ``` List SLO definitions for configuration review: ```bash elastic slos list-definitions ``` ## Understanding SLO Data ### Key Concepts - **SLI (Service Level Indicator)**: the metric being measured (e.g., error rate, latency) - **SLO target**: the goal percentage (e.g., 99.9% availability) - **Error budget**: the allowed amount of "bad" time/events (e.g., 0.1% of the window) - **Burn rate**: how fast the error budget is being consumed ### SLO Status Interpretation | Status | Meaning | Action | |--------|---------|--------| | Healthy | Within error budget | No action needed | | Degraded | Budget being consumed faster than expected | Monitor closely | | Breached | Error budget exhausted | Immediate investigation required | ## Querying SLO Data via ES|QL SLO data is stored in Elasticsearch indices. Query the rollup data: ```bash elastic es query 'FROM .slo-observability.sli-v3* | STATS good = SUM(slo.numerator), total = SUM(slo.denominator) BY slo.id, slo.name | EVAL sli = good / total * 100 | SORT sli ASC | LIMIT 20' ``` ## SLO Best Practices for SREs 1. **Start with user-facing SLOs** -- measure what customers experience 2. **Use error budgets for decision-making** -- if budget is healthy, ship faster; if degraded, focus on reliability 3. **Alert on burn rates, not thresholds** -- burn rate alerts catch sustained degradation without noisy spikes 4. **Review SLOs quarterly** -- adjust targets based on real-world performance