Products :: SPM :: HBase Performance Monitoring
"SPM is getting better and better every time I look at it. Keep
up the great work."
-- Logan Hardy, Operations Engineer, 33Across
SPM for HBase is a
proactive performance monitoring solution
available
in the Cloud (SaaS) as
well as
on Premises. SPM SaaS
requires
minimal setup. Within
minutes your DevOps and SysAdmins can start deriving valuable
insights from key operational metrics and optimizing performance
all without having to first set up, manage, or scale your own
monitoring systems and infrastructure. SPM has
integrated alerting and lets you view
real-time performance graphs and
charts, spot trends, narrow reports by time, by server, and by
other application-specific dimensions.
SPM can also graph
custom
metrics (e.g.
business
KPIs), which can be viewed along performance metrics on
custom dashboards.
(click to enlarge)
- Cluster and RegionServer read and write request counts
- Read, Write, and Sync min, max, and average latency
- Number of Stores opened on RegionServers
- StoreFile Index size and counts
- MemStore size
- Block cache: item count, hit cache and hit caching ratio, number of evictions, hits and misses, bytes free
- Compactions: compactionSizeNumOps, compactionTimeNumOps
- Compactions min, max, and average time
- Compactions min, max, and average size
- Compactions queue size
- Flush: flushSizeNumOps, flushTimeNumOps
- Flush min, max, and average size
- Flush min, max, and average time
- Splits: splitSizeNumOps, splitTimeNumOps
- Splits min, max, and average size
- Splits min, max, and average time
- HBase region count
- CPU usage
- RAM used/free
- System load
- Disk reads/writes
- Disk used/free
- Network interface traffic
- Swap IO
- JVM memory usage
- JVM garbage collection times and counts
- JVM thread counts
- Rich, near real-time performance graphs and charts
- Custom dashboards
- Application, server, and custom metrics all in one (+ integration with Coda Metrics)
- Integrated alerting on any metric
- Graph sharing end embedding
- Multi-user support with different user roles
- Powerful filtering: restrict graphs to one, multiple, or all servers at once, together or individually
- Filter data by multiple criteria at once: time, server, metric, etc.
- Compare performance of two time periods
- Multiple levels of report granularity
- Automatic graphing of extended time periods for easier long-term trend view
- No loss of precision for older performance data
- Easily switch between multiple monitored systems
- Minimal overhead
- Zero fixed costs - no software licences or server hardware to buy or maintain, just pay for what you use
- Simple - no complex software to install, configure, and maintain
- Easy and quick integration - run a single install script after you register and see performance charts in minutes
- Easy in, easy out - if you try SPM and don't like it, simply remove the small client-side piece that sends us your data