How does InterMapper store the data it collects from monitored devices?
InterMapper automatically saves data from probes that is designated ?Autorecord?. This function of the InterMapper server puts data into the Chart Data folder in InterMapper Settings. These include automatically saved data such as availability and response time. Probes can also be configured to autorecord specified data by default.
Users can collect and store additional data by configuring Strip Charts from data available in a device?s Status Window. By default, once the data collection is started for a set of data, it is collected continuously until deleted by the user.
Creating a strip chart from the Status Window of a device.
The storage of probe data in InterMapper can be changed using the data retention tools and policies discussed further on, allowing the user to control the type and amount of data stored by InterMapper to suit the needs of the specific situation.
To store data for retrieval and creating reports with, InterMapper provides a built-in PostgreSQL database. The Database is part of the InterMapper DataCenter, which is installed along with InterMapper on OS X and Windows (separate installation on Linux and Solaris).
The database can run on the local server, or be configured on a separate server as needed. Data may also be exported to a separate PostgreSQL database_if required.
The Database is started by the Reports Server, in the Edit > Server Settings tool.
Start data collection in the Reports Server tool in Server Settings.
The_Database collects information you specify and stores it for use in reports. The data is reduced (summarized) into collections for storage ? 5 minute samples, hourly_samples, and daily samples for each saved dataset._Your report history and granularity requirements for reports you generate will determine the type of data you store and how long you retain it in the database.
Data retention policies control what data is kept, where it is put, and how long it is saved before being discarded or overwritten. They control storage of data both in the ChartData directory and in the Database.
You can display, add, edit, or delete retention policies in the Server Settings->Retention Policies tool.
The Retention Policies tool in Server Settings.
Several policies come pre-configured in InterMapper, with self-explanatory names._Note above the ?None? policy. If the data retention policy for a chart is set to None, the chart will not show any data. If the policy is changed to Chart Only, data will appear on the chart and be saved to the ChartData directory. If changed to any of the other policies shown, data will also be written to the database, and retained for the length of time shown per policy and per data type.
The database takes the original (?raw?) data being collected by a probe and saves it, periodically reducing it (summarizing it) into 5 minute, hourly, and daily samples. Depending on the type of Reporting (how fine-grained the data needs to be) the original and summarized data can be made available for as long as needed._
If none of the existing policies shown in the Server Settings->Retention Policies tool match your needs, you can create a new one by clicking the plus sign (+) or edit an existing policy using the pencil icon._Once added, these retention policies become available for use immediately.
Retention policies can be set for the InterMapper server, individual maps, devices, and per dataset instance (e.g., link utilization on interface 5).
The Retention Policies tool edit screen.
When thinking about using InterMapper Reports, which is based on the InterMapper Database, a little planning is in order in order to make sure you have accounted for the disk space you will need.
The InterMapper Database requires a significant amount of disk space._It also requires a fair amount of free space on the disk in order to perform maintenance operations. Here are some guidelines.
1. Keep as little 'raw data' as possible._Reports barely uses the raw data, so unless you're using SQL to query the database, we might recommend the following:
_- Keep raw data for just a few days; not more than a week
_- Keep 5-minute samples for 1-3 months
_- Keep hourly samples for 6-12 months
_- Keep daily samples forever
Raw and 5-minute data tend to make up about 90% of the database's overall size, so the more you can limit those, the less disk space you'll need.
2. Once you've applied a good server-wide policy, wait for the database to go through another collection cycle, then use the 'Overview' page on its web interface to see estimated future disk usage._Add at least 50% to that value, and that's how much space you'll need to set aside._For example, if the estimated usage reaches 5GB and then levels off, you'll want to make about 8-10GB available for the database.
3. As you add more devices to your network, the database will grow beyond that initial estimate.
The InterMapper DataCenter Database Overview predicts data use.
It may be that maps have been laid out so that it would be convenient to have retention policies which apply to all or most of the devices on each map. To change the default setting for a map, use Edit-> Map Settings->Monitoring->Retention Policy.
Setting a default data retention policy for the server.
One of the retention policy settings, Inherit, causes the object (map or dataset) to use the policy of the parent object. So setting a map policy to Inherit will specify that the server-wide setting be used for the map.
Note that changing the Map retention policy will not change the retention policy of existing devices on the map, but only affect devices added to the map after the change._To change the setting for all existing devices on a map, change the map in list view, select all the devices, then, right-click/command-click_and choose Set Info->Set Data Retention.
List view for a map allows per-device modification of retention policies.
In Dataset View, devices and interfaces with chartable values appear in the left pane of the window. The chartable values for the selected devices and interfaces appear in the right pane, with type and retention policy for each dataset. Set or adjust retention policies for the values from this view.
Setting the database retention policy in Dataset View.
Caution: Enabling data retention for multiple parameters on many devices can adversely affect disk space.
For every poll cycle for any saved dataset on a given device, 32 bytes are written to an associated chart data file. For a 30-second poll interval, this translates to 60*2*32=1.632 Kb/hr, without consideration of the additional storage in the database (which varies with data size and length of retention for raw, 5-minute, hourly, and daily categories).
One other option exists for data collection and storage, in addition to saving it in the ChartData directory or in the Database ? configuring log files from charts. These are logs saved to the InterMapper Logs directory, containing tab-delimited data for one or more sets of data, recorded as it is received._
These data logs are not affected by the data retention policies in use for the data, and will be written whether or not the data is saved elsewhere. They can be retained forever or aged out after a period of time, at the discretion of the user who created them.
To create a log, just right-click/command-click on the data trace in a_chart and choose ?Create Log?.
Saving data into a log file.
Still have questions? We can help. Submit a case to Technical Support.