Doberman

Uptime Monitor for ScottyLabs

Launch

Install Docker first; can check for successful installation with docker --version

# run/stop
docker-compose up -d    # run
docker-compose down     # stop

# debugging
docker-compose logs                    # show all logs
docker-compose logs blackbox-exporter  # show logs for blackbox
docker-compose logs prometheus         # show logs for prometheus

Target for Prometheus is set to http://localhost:9090/. Go to Alerts panel on the top toolbar for alerts info, and Status > Target health panel for current endpoint statuses.

So far, there are two alerts: WebsiteDown (fires if website has been down for > 1 minute), and WebsiteSlow (fires if website takes > 5 seconds to respond). Note that http://httpstat.us/503 is expected to be down.

Note

Before running, make sure to create data/prometheus/queries.active inside of the root directory, otherwise Prometheus will fail. This is not the expected behavior, as Prometheus should bootstrap itself on first startup - not sure how to fix this, may have to do with chown nobody. See prometheus/prometheus#5976 for details.

TODO

• fix the aforementioned problem
• make alertmanager
• customize alerts
• figure out Grafana
• figure out what server to run prometheus on
• :p