Monitoring, Logging and Alerting
To install graphite on an AWS instance:
- Open incoming ports 80 and 8125/UDP
- Install docker and docker-graphite-statsd, then set the
STATSD_ENDPOINTvariable on metric-exporter to
- If username and passwords are set correctly, the web interface should show that data is coming in. If not, ssh on the box and run
sudo ngrep -d any -W byline port 8125.
To monitor Redis metrics, we use redis-statsd (just the script, not the container) running on the graphite instance.
We are currently in the process of switching to Prometheus as provided by Reliability Engineering. We have a folder in Grafana which you can sign in to with your GDS email. This works by the ‘dgu-prometheus’ service in PaaS scraping the
/metrics endpoint for our apps.
If you wish to see which metrics are available, you can use the following command to see what a specific app publishes:
curl -H "Authorization: Bearer $(cf app find-data-beta --guid)"