I deploy and configure the Grafana + Loki + Promtail observability stack for centralized log aggregation across your servers, containers, and Kubernetes workloads — so all your logs are queryable from one place, with dashboards and alerts included.
Loki stores logs compressed without full-text indexing, making it significantly more cost-efficient than ELK for most use cases while still providing powerful LogQL queries. I configure: Promtail agents on all target hosts; label schemes for filtering by service, environment, and host; application log parsing for common formats (nginx, JSON, syslog); log-based alert rules; and Grafana dashboards for error rate monitoring.
For Kubernetes environments I configure a DaemonSet Promtail deployment to collect pod logs from all nodes. I provision Grafana dashboards and alert rules as code (JSON/YAML) so they're reproducible and version-controlled.
If you already have Grafana running, I add Loki as a datasource. Metrics and logs can coexist in the same dashboards. Alert routing to Slack, PagerDuty, or other is included.