// the find
uschtwill/docker_monitoring_logging_alerting
Docker host and container monitoring, logging and alerting out of the box using cAdvisor, Prometheus, Grafana for monitoring, Elasticsearch, Kibana and Logstash for logging and elastalert and Alertmanager for alerting.
A docker-compose bundle wiring cAdvisor, node_exporter, Prometheus and Grafana for metrics, plus Filebeat, Logstash, Elasticsearch and Kibana for logs, with elastalert and Alertmanager handling alerts and an nginx+letsencrypt proxy for a 'secure' mode. Aimed at someone who wants a full Docker host monitoring/logging/alerting stack without assembling each piece themselves.
Pre-wired Grafana data sources and a dashboard ship as mounted volumes, so you get a working dashboard on first boot rather than building panels from scratch. The secure-mode setup script automates real infra work: subdomain routing, automatic SSL via the letsencrypt companion container, basic auth, and no manual port exposure. Any container opts into monitoring and log shipping through one convention (container_group label + gelf driver), which is a clean way to keep the compose files simple. It also links out to a companion repo of human-readable descriptions for cAdvisor/Prometheus metrics, filling a real documentation gap upstream.
Last commit is 2018, pinned to ELK 6.3 and that era's Prometheus/Grafana — a plain docker-compose up today is likely to break against current image tags and config formats, not just feel dated. Kibana index patterns and dashboards were removed after the 6.3 upgrade per the README, so the 'out-of-the-box' logging story now requires manual index setup. Setup/cleanup logic lives in untested shell scripts with a documented umask footgun (bad umask breaks Kibana's config permissions on first run), and there's no CI. storage/grafana/grafana.db and session files are committed directly to the repo, so cloning it hands you someone else's stale Grafana state instead of a clean install.