Discover CloudDB AF

Automate, plus See and Use your Stats in Whole New Ways!

Core Concept

  • Metric - The metric collects data and replicates it to the repository.
  • Gauge - The gauge analyzes the data and evaluates it against thresholds.
  • React - The reaction performs one or more corrective or supportive actions based on gauge results.


  • Process-oriented - Modular framework; dedicated processes performing set tasks.
  • Centralized - Centralized repository where it all can be managed. Can't be overridden on the target
  • Graphical UI - Easy management with an intuitive user interface.
  • DBMS_SCHEDULER - Framework actions triggered by the powerful Oracle job scheduler
  • Full Pl/SQL API - Fully scriptable and quick integration with configuration management such as Ansible, Puppet.

Blackout Modes

  • Full Blackout- Framework shuts down with the exception of processes needed to end blackout.
  • Silent Blackout - Framework remains active and collects data but will not trigger reactions.
  • Flexible Windows - Timed windows and blackout schedules allow for flexible blackout plans.

Example of Quick Benefits

  • Automatically add data files to your tablespaces when they became full and as long space is available.
  • Automatically resolve contentions by terminating offending sessions.
  • Produce detailed reports of contentions that can't be automatically resolved to enable quick diagnosis.
  • Easily analyze high process memory usage through provided break-down reports.
  • Automatically track interesting SQL that was observed to have caused issues in the past.
  • Collect information in a real-time when an issue happens, making the information available to you to analyze the problem later.
  • CloudDB AF writes important supplemental information with regards to your database health to the database alert log.


  • Quickly add custom metrics to collect additional data points of interest.
  • Easily add new gauges to introduce new checks on the collected data.
  • Write your own reactions in PL/SQL and quickly integrate them into the framework.

Amazing GUI

Centralized Dashboard

From the initial idea, CloudDB AF was designed to make the DBA more productive and cost-efficient. It provides prompt ROI through its automated & reactive nature. It supports the DBA by providing the required diagnostics automatically when a situation that requires intervention arises. The framework is highly customizable and easily extensible. Automation is the future, and that's exactly what CloudDB AF does best!

** ( This is not a complete list. It merely highlights some of the areas CloudDB AF can handle out of the box )

Endless possibilities


  • Archive Destinations - Status of archive log destinations
  • Archive Logs - Archive log production volume
  • ASM Disks - State of the ASM disks.
  • ASM Diskgroups -  State and space usage of ASM disk groups.
  • Blocked Sessions - React to blocked sessions.
  • Connection Pools - Expected number of sessions within a specific range. For middle-tier connection pools.
  • Database Heartbeats - Database availability.
  • Database Upgrades - Automatic handling of database upgrades and refresh of internal metrics.
  • Datafiles - Space usage and accessibility.
  • Extents -  Extent usage.
  • Idle Sessions -  React to idle sessions based on the client used, or origin, etc.
  • Latch Contentions - With optional high-speed sampling on x$ksuprlat, automatic report production.
  • Marked SQL - Mark SQL to be tracked.
  • Mutex Contentions - Optional high-speed sampling on x$mutex_sleeep, automatic report production.
  • Option Usage Alerts - React when a licensable option is used.
  • Session Memory Use -  Detailed PGA use reporting, down to Oracle kernel area layers.
  • Session Stats - Session statistics based on v$sesstat.
  • Session Waits -  v$session_wait with optional high-speed sampling on x$kslwt.
  • Tablespace - Space usage and availability. Automatic datafile addition.
  • Top SQL - Based on exec, fetch, ela, disk reads, etc.
  • Used Features - Maintains history of dba_feature_usage_statistics.
  • User Passwords - Can react before a password expires.

Flexible Analysis


  • Sending Emails - Traditional, but with added details whenever available.
  • Alert Log - Writing messages to the alert log to indicate a problem.
  • ZZtat Log - Everything is tracked there, but alerts can also be written there.
  • Custom Logfile -  Write to a custom log file.
  • PGA Reports - Light non-intrusive version and a super-detailed version that uses pga_detail_get.
  • Latch Reports - Detailed breakdown of latch gets.
  • Wait Reports - Detailed report of session waits.
  • Session Termination - Automatic termination of a session.
  • Datafile Addition - Automatic data file addition, as long as space is available. ASM only.
  • Marking SQLs -  Marked SQL is tracked and has its execution plan & statistics collected.
  • High-Speed Sampling -  x$kslwt, x$mutex_sleep and x$ksuprlat for now. More to come!
  • System State Dump - Never be faced with an SR that requests information you can't get any more.
  • Hanganalyze - If a certain percentage of sessions hang, collect hang analyze automatically.
  • Library Cache Dump - Trigger when extreme parsing activity is seen.
  • Row Cache Dump - Trigger when shared pool memory errors are seen.

** ( This is not a complete list. And of course you can define your own reactions )

Proactive Reactions

Reaction Chains

  • Add Datafile + Send Email - You can combine reactions at will by assigning them to reaction chains.
  • Mark SQL +High-Speed Sampling +Latch Report - Latch reports are far more detailed if high-speed sampling is performed before the report is taken. 
  • Mark SQL +PGA Report +Terminate Session - If a session was terminated because it jeopardized stability, you'll know exactly why.

** ( Reaction chains are fully customizable; any reaction can be fired as part of a chain)

Flexible and Functional

Built from the ground-up with security in mind

  • No connections to your production databases (in the default operation mode)
  • No agents (this also means there are no huge memory- and CPU-hogging processes running on your servers.
  • Only a single package resides in SYS (and that's optional, too!) - the entire framework runs as a normal unprivileged user.
  • Limited access is used whenever possible (CloudDB AF follows the principle of least privilege at its core) No ANY-privileges (one exception: CREATE ANY CONTEXT, since Oracle doesn't provide a CREATE CONTEXT privilege)

Unique PGA Handling

CloudDB AF maintains a table of Oracle kernel functions, their mnemonics, and a description.

When the PGA report gets its data from v$process_memory_detail, the NAME and HEAP_NAME columns would refer to Oracle kernel areas, such as kdbl or qmrb as per example.  Any normal monitoring tool ends there and you're then left to perhaps search My Oracle Support for any possible meaning of that NAME or HEAP NAME.

CloudDB AF powered by CloudDB AF automatically tries to match PGA areas to its kernel function library. This means you will immediately see right there in the report if e.g. the session is using so much memory because it's processing XML (qm*) or PL/SQL collections (pmu*).

Coming Soon with the first Patchset.

The EYE - The CloudDB AF Eye

  • Live monitoring as well as interactive alert handling
  • Secure: 2-factor authentication
  • Mobile UI: Handle alerts right on your phone.
  • Reactions triggered by you are forwarded to targets and executed live.
  • Get more information about the session that triggered the alert right onto your device.
  • Only needs a connection to the repository from your mobile.
  • Alerts related to sessions and tablespaces will be included in the first version of the eye.
View Details
Sold Out