Monitor all Master, Worker, Driver, and Executor metrics
- Correlate metrics, anomalies, alerts, events & logs
- Distributed Transaction Tracing with custom pointcuts
- Out of the box graphs and custom dashboards
- Filter metrics by server, time, etc.
- Built-in anomaly detection, threshold, and heartbeat alerts
The 1st Spark Monitor
Spark has a Master and Workers, Drivers and Executors, along with Spark Streaming. Which of them should you monitor? All of them, of course! SPM captures all Spark metrics and gives you performance monitoring charts out of the box. Setting up anomaly detection or threshold-based alerts on any combination of metrics and filters takes just a minute. Heartbeat alerts, enabled by default, notify you when any of your nodes goes down.
Full Stack Visibility
Spark doesn’t run alone. There are other apps around it. Perhaps you also run Kafka, HBase, or Cassandra? Perhaps Elasticsearch, Solr, or HDFS? SPM monitors them all, so you can have a single place to see your full stack. And not just metrics, but also alerts, anomalies, arbitrary events, and even logs when you ship them to Logsene so you can find out not only WHEN something happened, but also WHAT and WHY.