Roy Russo
Roy Russo
We are currently keying off of cluster_name. Some users have separate clusters with the same name, so the solution is to use the cluster_uuid along with the name to differentiate.
Removing develop branch and tagging master is more in-line with GitHub flow: https://softwareengineering.stackexchange.com/questions/312022/the-trend-of-the-develop-branch-going-away
Per alert rule: - allow for polling interval. - push type (slack, email, twilio, etc...) ... these will need config screens of their own?
- for now, a simple table view per cluster will work.
- Alert rules triggered should save a log within the cluster. - explore using something like `.elastichq_alerts_log` for storing records of past alerts.
- can run on a predetermined schedule - can monitor for metrics per cluster: low disk space, high heap ... threshold would vary per cluster.
http://python-rq.org/
- store alerts settings under the same cluster it's alerting on. - explore a new index `.elastichq_alerts_config` for storing the settings