Version 4.5

Changes since the 4.5 main release

Improvements to Traplogger for new storage devices.

Exchange 2010 monitoring

Allow duplicate nodes for monitoring of websites hosted on virtual servers.

Fixes to SMS messaging

NEW polling for ESXi5 

Front End Changes

Improved X-browser support for IE8, IE9, Firefox 6 & Safari 5. Still some issues with Google Chrome, but most features work OK.

On tablet devices (e.g. iPad), single-click selection is automatically enabled.

Where tables of data are displayed (Event/Alert Logs, Table View, Open Events), two-tone "swim lanes" are now used to make reading easier.

Open panels with graphs now automatically update every minute. This is most noticeable if the time period is set to "Last Hour".

The Interfaces panels have been redesigned. In the main panel, only interfaces that have been selected for polling are displayed. These are chosen from a new Interfaces-Configure sub panel. From this panel you can enable two new features:

• The Interface Description can now be simply switched between the two SNMP-derived names (Alias or Description), where these are available. You can also configure custom descriptions.
• It is possible to override the SNMP-derived interface bandwidths with manually-configured values. This is useful for asymmetric lines (e.g. ADSL) where SNMP is only able to give one of either input or output bandwidth (& not both) or in cases where a hypervisor fails to correctly notify the guest system of the physical interface values, for example. In both these cases usage figures or greater that 100% have sometimes been displayed historically.

Setting of interface-bandwidth thresholds has also been moved to a new Interfaces -Thresholds panel.

For SNMP-monitored Windows and Linux-based systems, a new [Software] button is shown. Clicking this will display the list of the currently-installed software that has been detected. If the [Software] button is not displayed, the node may need to be reset - see below.

Options for resetting the polling for one, or a set of nodes, have been simplified. Clicking the [Reset Polling] button at the bottom of the Configure panel for a individual node will now display a new Reset-Polling panel with 5 options:

• Update - re-reads basic SNMP information from the node's SNMP tables, including the network interfaces. It does not change any polling parameters (unless interfaces have been added, removed or renamed). This data is displayed by clicking the [SNMP Info] button at the bottom of the node Status panel.
• Safe - re-reads all the data above, but additionally Mutiny reads the node asset properties (such as OS, disk capacities, serial numbers, installed software etc.) where these are available. These are displayed by clicking the [Properties] or [Software] buttons ate the bottom of the Status panel. No polling settings are altered.
• Add new features - This performs the same update as "Safe", but additionally, it will try and add in the monitoring of any node properties that Mutiny thinks should be monitored, but are not. Mostly this will be new Agents and Adapters that have been added to Mutiny since the previous upgrade, but it will also include any new disks and critical processes.
• Recalculate Thresholds - this is pretty well a full reset of the monitored properties and thresholds of the node, back to the values that would be set if the node was deleted and re-added. The difference, however, is that the reset will preserve historical data where as delete & re-add will not.
• Custom - resets the node using the custom settings that have defined in [Default Settings] => [Reset Options] in the Node Manager View.

The above same reset options are available to operate on multiple nodes or Views using [Update Nodes] from Node Manager.

NOTE, to get the best out of Mutiny, it is recommended that following the upgrade to Mutiny Version 4.5, the entire monitored estate is updated from Node Manager by clicking [Update Nodes] and then selecting "Add New Features" and then the "All Nodes[S]" smart View and clicking [Apply]. This will take several minutes depending on the number of nodes being monitored. Beware that Mutiny may add back in some Agents that you may previously have manually removed from some nodes. One example is the "HP PSU" Agent on severs where only one power supply is connected.

Support has been added for Mutiny Restorepoint and the new Mutiny external Email/SMS gateway. If these add-on devices have been purchased, simply adding the appropriate IP addresses in the [Admin] => [System Configuration] page is all that is required to enable them in Mutiny.

It is now possible to set any Contact as a "Default Contact". This will cause ALL Email Alerts, either Critical & Warning - or just Critical - to be sent to the contact without need for further Event/Alert configuration. This is useful where you would like one, or more, Email accounts to keep a log of all incidents detected by Mutiny - one example might be a Help Desk or a Service-Desk system.

The default Email Alert template has now been changed to be the "short-format" template. This change is automatic, so if you wish to continue using the old -style template for any Contact, you should select "long_format".

Both of the above are selected from the [Admin] => [Contact Administration] pages.

It is possible to test Mutiny's Alerting mechanisms on either a daily or a weekly basis by choosing one of the new "Weekly Alerting Test" or "Daily Alerting Test" Agents on a chosen node. The test runs at 08:00.

The ordering of nodes in the Smart Views (e.g. All Nodes) has been changed to be predominately alphabetical by node name.

Some additional icons have been created - most notably NetApp and other network-attached storage devices.

The Front-end Web server has been upgraded to Tomcat 6.

Monitoring

Where available, 64 bit interface counters are chosen for monitoring network interfaces over the older SNMP v1 counters.

There have been many improvements to the way that Mutiny detects and displays SNMP traps and informs that it receives. The SNMP trap Agent is now automatically added to many type of device (it can still be manually added or removed if desired) and dedicated Trap Interpreters have now been added for some specific types of device. In summary the changes are:

• Most SNMP traps should now be decoded and useful information should be presented in the interface, Event Logs and Alerts.
• Where the information is not useful, for example if the Status is wrong or misleading, or the data is not clearly presented, the interpretation can be quickly altered with a simple patch file. Please contact support@mutiny.com if you are receiving such traps.
• Dedicated support added for traps from HP EVA storage systems
• Dedicated support added for traps from VMware ESXi
• Dedicated support and re-writing for Mutiny Restorepoint traps.

Bugs with the polling of UPS devices and HP Blade Chassis enclosures have been addressed.

An Application Engine has been added which will make it simple to create Apps to monitor simple non-SNMP properties or to manipulate polled data. We are planning to release new Mutiny Apps on a monthly basis.

Behind the Scenes

The process of writing the historical information into the RRD (Round-Robin Database) files has been completely revised and a new caching methodology has been implemented. The aim has been to greatly reduce the loading on the IO performance of Mutiny. In some cases, users have reported unacceptably long (greater than 2 minutes) polling cycles when polling greater that 500 nodes with the delay being caused by writing data to disk. This should now be much improved and polling cycles should be closed to the desired per-minute schedule.

The performance of the backup process has been improved and this will now have a lesser impact on the polling.

Node properties are now written daily in a ".CSV" format file named "NodeProperties.csv" in the "/var/Mutiny/" directory. This file can be downloaded and read into Microsoft Excel (or similar) by setting a "Backup user" password in the [Admin] => [Backup/Restore] page.

The NetFlow collector agent has been improved and various watchdogs have been added to monitor its performance. It is recommended that an additional Mutiny appliance, running on dedicated hardware, is used as a NetFlow collector when receiving NetFlow from any LAN-based switches and/or more that 5 WAN routers. The main Mutiny system can then be used as the NetFlow viewer with data residing on the dedicated collector. This configuration removes the risk of heavy NetFlow data damaging the performance of the main Mutiny monitor or any other VMs that are hosted on the same physical server.

Licence restrictions have been removed for the old Windows Agents to monitor Exchange, SQLserver, IIS and the Windows Event logs.

There have been some changes to the way in which expired Mutiny licenses are handled and, in most cases, additional warnings will be displayed before polling is disabled.