​
Loki + Promtail

Loki Port: 3100 (internal) Config: /opt/coderz/configs/loki/loki-config.yml Promtail Config: /opt/coderz/configs/promtail/promtail-config.yml Developed by: Grafana Labs Loki is a horizontally-scalable log aggregation system designed to work natively with Grafana. Unlike Elasticsearch, Loki indexes only the labels (metadata) of logs, not the full content — making it much cheaper in terms of storage and CPU.

​
How Promtail Collects Logs

Promtail runs on the same host as the Docker daemon. It:
  1. Reads Docker container log files from /var/lib/docker/containers/
  2. Attaches labels (container name, image, compose service)
  3. Ships logs to Loki in real-time

​
Log Labels

Every log line in Loki is tagged with:
LabelExample
container_namecoderz-grafana
compose_servicegrafana
image_namegrafana/grafana:latest
filename/var/lib/docker/containers/abc123/abc123-json.log

​
Querying Logs with LogQL

LogQL is Loki’s query language. It works like grep but with label selectors. 
# All logs from the .NET API
{container_name="coderz-dotnet-api"}

# Only error lines from the .NET API
{container_name="coderz-dotnet-api"} |= "error"

# Logs containing "500" from any container
{compose_service=~".+"} |= "500"

# Count errors per minute
rate({container_name="coderz-dotnet-api"} |= "error" [1m])

# Parse JSON logs and filter by status
{container_name="coderz-dotnet-api"} | json | status >= 500

​
Grafana Integration

Loki is pre-configured as a Grafana datasource. In Grafana, use it in:
  • Container Logs dashboard — live log viewer for all containers
  • Prefect Flows dashboard — flow run logs and completion events
  • .NET API Full Stack dashboard — live API request log table
To explore logs manually:
  1. Open Grafana → Explore
  2. Select datasource: Loki
  3. Enter a LogQL query

​
Loki vs Elasticsearch

FeatureLokiElasticsearch
Index approachLabels only (metadata)Full-text index
Storage costLowHigh
Query languageLogQLKQL / Lucene
Best forLabel-based filteringFull-text search
Grafana nativeYesNeeds plugin
Resource usageLow (~50MB RAM)High (~1–4GB RAM)