The SPM Change Log: new SPM features, enhancements, and bug fixes.
Features & Improvements
- Added ability to run SPM Monitor as a standalone process
- Made graph URLs sharable among SPM application users, including preservation of time and filter selection in URLs
- Alert emails now contain the link to the specific graph that shows metric which triggered an alert
- Made SPM client installation much simpler for Solr users Solr
- Made Sender be nicer with IO via ionice
- Made Sender differentiate between errors vs. disabled/deleted apps
- Simplified Collectd config
- Reworked SPM client installation and reworked how SPM Monitor and SPM Sender talk to each other; made both Monitor and Sender lean and bootstrapped
- A number of behind the scenes improvements and fixes
Bugs
- Fixed blank search latency graph - thanks David Smiley! Solr
Features & Improvements
- Added paid plans
- Added ElasticSearch HTTP metrics ElasticSearch
- Added ElasticSearch TCP metrics ElasticSearch
- Made graphs truly show the last 1 hour instead of rounding to the next full hour
Bugs
- Differentiate between a dead server and a dead application to avoid false application heartebeat alerts
Features & Improvements
- Deployed real-time, in-memory data aggregators
- Switched to real-time metric graphs
- Simplified installation
- Deployed faster Receivers
Bugs
- Fixed blank Disk Space and Network IO graphs caused by erroneous directive in collectd.conf. Agent upgrade is required to fix this.
- Fixed a bug with false alerts
Features & Improvements
- Made JMX metric collection handle MBeans with multiple prefixes Solr
- Simplified app registration
- Simplified installation
- Made Receivers Highly Available
- Made Alerts subsystem more robust by persisting information across restarts
Bugs
- Small bugs not worth mentioning :)
Features & Improvements
- Added SPM for Java - support for monitoring of any Java application
- Added a system overview dashboard
- Multi-user support - made it possible to invite others and give them regular user or admin roles; added a notion of an application owner
- Changed graphs to display time in user-specific timezones
- Enabled application re-activation
- Added ability to send performance metrics via an HTTP proxy
- Alert creation UI cleaup
- Improved signup flow and application addition flow
- Numerous backend improvements - made the Alerts subsystem Highly Available, changed Alert rule storage, improved client installer, etc.
Bugs
- Small bugs not worth mentioning :)
Features & Improvements
- Alerts - heartbeat and performance alerts
- Email subscriptions to performance reports - daily, weekly, monthly...
- SPM Agent version tracking and notifications about outdated Agents
- Added more detailed Java Garbage Collection reports for Java 7 users
- Added the ability to filter Java Garbage Collection metrics by type
- Added report for the number of file handles opened by the JVM
- Added more detailed CPU reports
- Added more detailed Memory reports
- Improved Disk IO report: writes are bars, reads are areas, stacking in on by default
- Improved Disk Usage report to better show a percentage of free/used disk space
- Reworked database queries for speed: faster logging in, retrieval of existing systems, etc.
- Added SPM version to the UI ;)
- Simplified SPM install and reduced SPM client download size by 20%
- Added ability to apply arithmetic expressions to multiple metrics to derive and graph new measures (internal, server-side for now)
- SearchComponent performance breakdown with ability to turn it on/off on the fly Solr
- Added circles for Maxs Docs metric on Index report Solr
- Added MemStore upper and lower flush limits to MemStore Flush reports HBase
- Show writes blocking memstore size threshold on memstore size chart HBase
- Switched to using dots to represent non-data nodes on ES Nodes graph ElasticSearch
Bugs
- Fixed bar chart alignment with graph grid
- Fixed redirecting to the last used system after login
- Fixed broken layout in PDFs used in email subscriptions
- Stopped reporting reserved disk space as free