Loki + OpenTelemetry Learning Roadmap

Goal: From zero log-aggregation experience to production Loki operations, leveraging existing Prometheus knowledge

Learning Phases

Phase 1: Loki Fundamentals (1-2 days)

•Loki vs ELK architecture — why label-only indexing is cheaper

•Labels and streams — the same mental model as Prometheus label sets

•LogQL basics — stream selectors, line filters, | json/| logfmt parsing

•Promtail deployment and automatic Kubernetes pod-label attachment

Phase 2: LogQL and Alerting (2-3 days)

•Metrics-from-logs queries (rate, sum by)

•Advanced LogQL — unwrap, quantile_over_time, deriving metrics from unstructured fields

•Grafana panel and dashboard building from LogQL queries

•Ruler-based alerting directly on log content, not just derived metrics

Phase 3: Production Storage and Retention (2-3 days)

•Chunk vs. index storage model — what each actually stores and why the distinction matters for cost and query performance

•Retention configuration — retention_period + compactor.retention_enabled together, not either alone

•Cardinality safety nets — max_streams_per_user and what happens when it's hit

Phase 4: Scaling and Advanced Operations (1 week)

•Distributed/microservices installation mode vs. the loki-stack bundled chart

•Independently scaling ingestion vs. query load

•Multi-tenancy configuration

•Alerting rules via the Ruler at scale

Phase 5: Certification and Interview Readiness (ongoing)

•See certification.md for Grafana Labs' current Loki-relevant certification options — treat specifics as (needs verification — recheck against current source)

•Practice explaining the chunk-vs-index tradeoff and the ELK-cost comparison out loud — both are recurring interview topics for this guide

Job Roles This Enables

•Observability/Monitoring Engineer

•Site Reliability Engineer

•DevOps Engineer (logging-focused)

•Platform Engineer maintaining a Grafana-stack-based internal platform

Target Certifications

See certification.md in this guide.

How to Use This Roadmap

1.If you already know Prometheus/PromQL, Phase 1 will move quickly — treat it as a mapping exercise, not new fundamentals

2.This technology doesn't have a separate Labs tab — the Fundamentals/Intermediate/Advanced code blocks and the Installation section's verification steps are the hands-on material

3.Build the portfolio projects in projects.md after Phase 3, once retention/storage concepts are solid enough to configure correctly rather than by trial and error

4.Use the Troubleshooting section's real scenarios as a study aid, not just a reference — working through "why did this fail" builds the same diagnostic instinct an interviewer is testing for

5.Revisit Phase 2's Ruler-based alerting after Phase 3 — retention and storage decisions affect how far back an alert rule can meaningfully query

Prerequisites

See prerequisites.md in this section for what you should know before starting.

Revision Notes

Total time: 2-3 weeks (part-time), faster with strong prior Prometheus experience
The chunk-vs-index distinction and the retention_period + compactor.retention_enabled
pairing are the two concepts most likely to trip up someone new to Loki specifically —
budget extra review time there even if the rest of the roadmap feels fast

Loki + OpenTelemetry — Roadmap