Pandora: Documentation en: Architecture
- 1 The Pandora FMS Architecture
- 1.1 Pandora FMS Servers
- 1.1.1 The Data Server
- 1.1.2 The Network Server
- 1.1.3 The SNMP Server (also known as the SNMP Trap Console)
- 1.1.4 The WMI Server
- 1.1.5 The Recon Server
- 1.1.6 The Plugin Server
- 1.1.7 The Prediction Server
- 1.1.8 The web server (Goliath)
- 1.1.9 The Export Server
- 1.1.10 The Inventory Server
- 1.1.11 The Event Correlation Server
- 1.1.12 The Enterprise Network Server for SNMP and ICMP
- 1.1.13 Satellite server
- 1.2 The Pandora FMS Network Console
- 1.3 The Pandora FMS Database
- 1.4 The Software Agents of Pandora FMS
- 1.5 Topologies, Schemes and Monitoring Models
- 1.1 Pandora FMS Servers
1 The Pandora FMS Architecture
This chapter provides a general description of Pandora’s components, the way they relate to each other and how to use the Pandora FMS architecture to meet different challenges regarding the topology of its infrastructure.
Pandora FMS is modular and decentralized. The most important component is the MySQL database, where everything is stored. Each component of Pandora FMS can be replicated and works in a pure HA environment, be it passive, active or in a clustered environment (Active/Active with load balancing). There are also descriptions for methods to setup a high availability SQL backend.
Pandora FMS consists of several elements - among them the Servers, which are in charge of collecting and processing data. The server also inputs the collected and processed data into the database. The console is the part in charge of displaying the data present in the database and of interacting with the end user. The software agents are the applications which run on the monitored systems (usually servers), collect the information and send it to the Pandora FMS server.
1.1 Pandora FMS Servers
Under Pandora FMS, there are twelve different servers in total, specialized in and responsible for the various tasks necessary to make Pandora what it is today. The servers are integrated into a single application under the general name of 'Pandora Server' which is a multi-threaded application (multi-processing) that executes in sub-processes (threads) each one being different to the instances or to the specialized servers of Pandora FMS.
The Pandora FMS Server is the element in charge of performing the pertinent checks. It verifies and changes them according to their results. It is also in charge of firing the alerts established by the Pandora server administrator to notify those monitoring the various target systems of events.
The Pandora FMS Data Server can work with high availability and / or load balancing. In very large architecture it is possible to employ several servers simultaneously to handle large volumes of information, and may be distributed by geographic or functional zones.
The Pandora FMS Server is always working and verifying if any monitored elements are experiencing any difficulties and can take appropriate action(s) if those are defined as alerts. When faced with a problem, it executes the response defined in the alert, such as sending an SMS, an email or activating the execution of a script.
There can be simultaneous servers, one of them being a main server and the rest of the servers being slaves. Although there is a master and a slave server relationship, they work simultaneously. The difference between them is that when one of the servers goes off-line, the other Pandora server continues to function, providing transparent and redundant Pandora services.
The server receiving the data file from the agent, or processing the information (if this is of the remote type) is the one which fires the associated alerts after the data is processed.
Pandora FMS automatically manages each server's status, load levels and other parameters. The user can monitor the state of each server by the server's status section of the web console.
1.1.1 The Data Server
The Pandora FMS Data Server processes the information sent by the Software agents. The software agents send XML data to the server in different ways (FTP, SSH or Tentacle) and the server periodically determines whether there are any data files awaiting processing or not. This process utilizes a disk directory as a bonding medium for the elements to be processed.
It's possible to install different data servers on different systems or on the same host (which will be different virtual servers). Several servers may work together in very large environments, making the best use of available hardware (e.g. multiple CPU environments).
The Data Server (like the rest of the servers) accesses the Pandora FMS database, which communicates with the web server, and contains the processed data packets. The server executes as a daemon or service and processes the gathered packets into its file system. In spite of its simplicity and its modest use of resources, the data server is one of the critical elements of the system as it processes all of the agents' information, and generates system alerts and events according to that data. The data server solely works by the XML data from the software agents without conducting any remote verification.
1.1.2 The Network Server
The Network Server executes remote monitoring tasks through the network: ICMP tests (ping, latencies) TCP and SNMP requests. If an agent is assigned to a server, it's always assigned to a web server, not to a data server. The reason for that is that it's very important that the engines executing the Web servers have 'network visibility' to be able to carry out the monitoring tasks assigned to them and making it possible for a server to connect to a defined network. For example, if we create a module for a ping verification on '192.168.1.1' and the agent or module is assigned to a server on '192.168.2.0/24' without network access to '192.168.1.0/24' we'll always get a 'down' reply, because it's unable to establish contact.
1.1.3 The SNMP Server (also known as the SNMP Trap Console)
The SNMP Server utilizes the standard daemon of the traps collection system called 'snmptrapd'. This daemon receives SNMP traps and the Pandora FMS Server processes and stores them in the database. While processing and analyzing them, it can also fire the designated alerts on the SNMP Pandora FMS console.
1.1.4 The WMI Server
WMI is a Microsoft standard to obtain information from a Windows-based operating system and Microsoft Windows environmental applications. Pandora FMS has a dedicated server to conduct native centralized WMI calls. Thanks to that server it's possible to collect data from Windows systems remotely and without the need for an agent.
1.1.5 The Recon Server
The Recognition Server (or 'Recon Server') is used to explore the network regularly and to detect new systems in operation. The Recon Server is also able to assign a monitoring template to recently detected systems and to apply the modules automatically by default, as defined by that template, so they can immediately be used to monitor the new system. By using the applications of the NMAP, xprobe and traceroute systems, it's also capable of identifying systems by their operating system, based on the opened ports and to establish the network's topology, guided by the systems it already knows.
1.1.6 The Plugin Server
The Plugin Server executes complex remote user tests conducted in any language, is integrated in the Pandora FMS interface and centrally managed. This allows the advanced user to define their own complex, self-developed tests and to integrate them into Pandora FMS.
1.1.7 The Prediction Server
The Prediction Server is a small component of artificial intelligence that implements a statistically predicted data set which is based on past data with a scope of up to 30 days in four temporal references. This allows us to predict the value of a data item in 10 to 15 minute intervals, making an educated assumption whether a particular data set presents an anomaly. At present, it's based on its past historical performance. You basically won't have to construct a dynamic baseline by a weekly profile. This server also manages the service monitoring calculations (BPM) from Pandora FMS versions 5 and above.
1.1.8 The web server (Goliath)
The Web Server (Goliath), the Export Server, the Inventory Server, the Event Correlation Server and the Enterprise Network Server are all features which are only available on the Enterprise Version of Pandora FMS.
The Web Server is used to carry out transactional web monitoring. It conducts synthetic network testing, which means complete network testing, including: User identification process, parameters for data transfer, contents verification, menu navigation, etc. It's basically intended for verification tests and to obtain latencies (in seconds) of the complete network navigation experience (including the resources linked to the page, like images, full texts, etc).
1.1.9 The Export Server
The Export Server of Pandora FMS permits the transfer of data from the monitored dispositive of a Pandora FMS installation to another, making it possible to facilitate replication of any data. This is particularly useful for large deployments with several Pandora FMS installations, for example, if we want to have some critical information centralized in only one of them.
1.1.10 The Inventory Server
The Inventory Server obtains and visualizes inventory information on the systems: Installed software and patches, hardware, memory, hard disks, services running on the system, etc. It's able to obtain this information both locally and remotely from the software agents.
1.1.11 The Event Correlation Server
The Event Correlation Server can be used to correlate events and generate alerts. This is a special server that does not monitor, and like the others, can be specified in the configuration for its start up. Compared to the rest, this server does not make use of the threads configuration or high availability.
1.1.12 The Enterprise Network Server for SNMP and ICMP
There are two additional servers using advanced strategies to process ICMP (ping) and SNMP (polling) requests in a way that offers superior performance to the open-source version, in exchange for some delicate requests (especially SNMP) as they work with previously validated OID’s by the opened server.
They use low level binary tools to access the TCP/IP system of the server by conducting block surveillance in a much more efficient way.
1.1.13 Satellite server
The Satellite Server is a special server, a hybrid between an agent and a server, it does remote monitoring, it also detects new systems, like a recon server, and executes remote plugins. It does not require a database connection to the central database. Instead, it sends XML data to the data server. You can read more about this later, in the specific section about Distributed monitoring with Satellite server
1.2 The Pandora FMS Network Console
This is the Pandora FMS User Interface. This administration and operations console allows the user to control the status of the agents, access statistical information, generate graphics and tables of data, as well as managing incidences with its integrated system serving different users with different privileges. It is also capable of producing reports and defining new modules, agents and alerts as well as creating other user profiles - and all of that is centralized from the interface.
The Web console is programmed in PHP and doesn't require the end user to install any additional software, not Java or ActiveX. Graphics are available, however, in Flash. To see them in that format you'll need the Flash application on your browser, which can be accessed from any modern platform which supports HTML and CSS. We recommend Firefox 2.x or IE 7.x. The user experience with older browsers like IE6 is very poor, because most of the features implemented in the Pandora FMS 3.0 console are unusable (and therefore hidden) if you're trying to use them.
The network console is able to operate and supervise multiple servers. This means you may add as many web consoles as you want, either for load distribution or to facilitate access due to logistical problems (large networks, several groups with different users, geographical and administrative differences, etc.) It's a prerogative to have access to the data storage center where Pandora FMS compiles everything. That means: Access to the database and in case of the Enterprise Version, synchronized access to the agent's repository of configurations via NFS.
1.3 The Pandora FMS Database
Pandora FMS utilizes a MySQL database. Pandora FMS hosts an asynchronous database with all the received information, performing a temporary cohesion and normalizing all the data received from several sources. Each agent's data module generates a data entry for each packet, which means a real production system can have a scope of ten million units of data or atoms of information.
This data is automatically managed from Pandora FMS which carries out a periodical and automated maintenance service of the database, allowing Pandora FMS to do without any manually assisted database administration system, be it by an operator or an administrator. This is done by periodically purging data after a certain time period (90 days in a preset option) as well as compacting the data which exceeds a predetermined and configurable number of days (the default value is '30 days').
1.4 The Software Agents of Pandora FMS
When we refer to an agent in Pandora FMS, we're describing three essential components in the collection of data:
- The master-agent, Pandora FMS.
- The software agent (software application, Pandora FMS agent running in a machine).
- The physical agent (hardware).
1.4.1 The Agent
The Pandora FMS agent itself is basically an organizational element, created with Pandora FMS Web Console and associated with a group of modules (seen as individual monitoring elements). This agent can also (optionally) have one or more IP addresses associated to it.
The agent can have associated remote modules, which would have been obtained by web servers, WMIs, plug ins, etc.
- Verification of whether the engine is connected or on line (PING).
- Verification of whether a given port is opened or closed.
- Verification of whether a network entity, hosted on a specific port of the hardware, is responding correctly.
- Verification of whether a network entity, hosted on a specific point of the hardware, has the desired content.
- Hardware verification(s) by SNMP (ascertaining the MIB).
- Latency time verification between the node and the Pandora FMS servers.
The agent can also have local modules associated with it. Local modules are those defined in the software agent's configuration which are also required to be defined within the network console's agent. If the agent is in 'auto-learning mode' (the default setting), these local modules are created automatically in the web console when a packet of data arrives from the agent for the first time. Therefore, an 'Agent' may contain modules of both remote and local types. The remote-type modules are executed by the servers obtaining information remotely (prediction included) and the local modules are obtained by the Data Server.
1.4.2 The Software Agent
A software agent installed on a remote node is completely different from the one on the Pandora server or within Pandora's network console. The software agent gathers information local to the node from the engine where it's executed, gathering information on the node by commands.
Pandora FMS Agents can be practically developed in any language as long as it meets the conditions of the data exchange API from the Pandora FMS Data Server (defined by the XML data exchange). The window agents operate in a free environment for C++ and employ the same interface and modularity as the UNIX agents, however coming with several characteristics of their own.
These scripts are built from sub modules and each one of them collects a portion of information.
Each agent collects several portions of information. These are compiled into one packet and stored in a single file named 'data package'.
The copy process of the data packet from agent to server is synchronously executed on a regular basis. That is, within regular intervals -defined by the agent- which can be modified in order not to clutter the data base with superfluous information, overload the web server, or to become detrimental to the system's performance.
The interval can be defined up to 300 seconds, which is the decimal equivalent of 5 minutes. Lesser values to 100 (seconds) are not recommended, as they can negatively affect the host-system's performance. Such a low polling time can overload the database and the central processing system.
It's important to remember that Pandora FMS is not a 'real time system', but a general monitoring system for systems and applications in environments where real time is not a critical factor. It may be adapted to operate in environments with response times of 3 to 5 seconds.
With either SSH or Tentacle, the process can be made secure, given that passwords don't travel through the network nor do they contain unencrypted confidential data, assuring the confidentiality, integrity and authentication of the connections between agent and server. The code-generating process and also the Tentacle protocol are detailed in the documentation on the installation and configuration of the agents and servers to be able to carry out the SCP (SSH) transfer automatically.
The transfer may also be conducted by FTP or any other file transfer protocols. However, we chose the Tentacle Protocol due to its security, user friendliness and the numerous options this system provides.
Please check the documentation annex regarding configuration of the transfer protocols.
Pandora FMS agents are designed to be executed from the agent they collect the data from, although the agents can also collect information stored in accessible engines from the host they are installed on. These are known as 'satellite agents'.
It is also feasible to configure a node in such a way to be able to bear several Pandora FMS agents simultaneously. This predicament is quite rare. It occurs if we have e.g. a software and a satellite agent. The standard software agent monitors the engine where it's executed while the installed satellite agents (there can be several) are monitoring remote systems by Telnet, SNMP or other proprietary protocols.
1.4.3 The XML Data File
The data file has the following syntax:
<host number>.<serial number>.data
This data file has an XML structure. It takes its name from a combination of the host names where the agent is located and a serial number, which differs in each packet, and the extension ".data" which indicates that this is a data packet.
<host number>.<serial number>.checksum
The ".data" file is the file which holds the data. The verification file with the ".checksum" extension contains an MD5 hash of the data file. These allow us to perform a final verification to ensure the data hasn't been altered in any way before being processed.
The XML data file the agent generates is at the heart of Pandora FMS. It contains a data packet along with the information gathered by the agent. This data packet has a compact design: Light and flexible that allows any user to use Pandora FMS agents or to generate information to be processed in Pandora FMS by other methods. The data file is an XML file similar to the following:
<agent data os_name=”SunOS” os_version=”5.8” timestamp=”300” agent_name=”pdges01” version=”1.0”> <module> <name>FTP Daemon</name> <type>generic_proc</type> <data>0</data> </module> <module> <name>DiskFree</name> <type>generic_data</type> <data>5200000</data> </module> <module> <name>UsersConnected</name> <type>generic_data_inc</type> <data>119</data> </module> <module> <name>LastLogin</name> <type>generic_data_string</type> <data>slerena</data> </module> </agent_data>
1.4.4 The Physical Agent
Pandora FMS has a physical agent mounted on an ASUS and an Arduino automaton. This tandem, along with the connected sensors facilitates the monitoring of the following environmental features:
- Ambient lighting
Because of the electrical nature of the sensors they are easily calibrated. Their values are also able to be processed by Pandora FMS without any difficulty. The fact that the sensor is a wireless router opens up an entire world of possibilities to this type of sensors already present in the CPDs of some companies.
1.5 Topologies, Schemes and Monitoring Models
There are different models to address the monitoring process, both local and remote. We enumerate the following common examples for different topologies in order to familiarize the reader with the possible problems and the solutions Pandora has to offer. Each of the solutions is described in successive chapters.
1.5.1 Accessible Networks
This is the norm in small, simple networks but also in the very centralized and well organized ones. This one is the easiest model to implement.
- Network access for centralized remote monitoring. It implies that we can access every node from Pandora’s server to probe remotely.
- Network access for agent based monitoring. In this network, we're able to reach Pandora’s server from the agents installed on the monitoring engine.
1.5.2 Limited-Access Networks
- Remote Network: This is an unreachable network for remote testing by Pandora FMS. We are using a software agent as a remote gatherer to test other systems on that end. We call these 'operating modes': A 'satellite agent mode' (when all testing is carried out within the same agent) and a 'broker agent mode' (if it impersonates several agents but all tests are actually carried out in the same physical engine).
Deployment model for remote and inaccessible networks in Broker Mode
- Software agents without access to a Pandora FMS Server. In this case, we're going to use the proxy characteristics of the software agents, allowing those agents without direct access to the server to use another agent with direct access to connect and facilitate the transactions.
Deployment model for remote networks by using the Proxy Agent Mode
- The need to conduct remote server monitoring for different networks: In this situation we're going to mount several different servers of Pandora FMS. Connected to the same database, one server is going to execute a battery of predefined tests and a different one by another. Both servers operate within the same environment which are being managed from the console simultaneously.
1.5.3 Special Organizational Characteristics
- The need to monitor several headquarters by monitoring equipment and different configurations. In this case, we're using an export to duplicate a part of the monitoring in an independent environment segregated from Pandora FMS.
Hierarchical export model along with an Export Server
- Duality of Reporting: We can configure additional agents to support two different Pandora FMS Servers, although only one will be able to manage it.
- Fragmented Management: It's pretty useful if you're required to delegate the administration of part of the equipment to different personnel with different access levels. This is more of a management issue rather than an architectural problem. It can be resolved by the assigned permissions within the management policies.
1.5.4 Large Environments
- A Large-Volume Network, consisting of thousands of network testing processes which we distribute within different 'remote monitoring probes'. Given their large numbers (over 50,000) we can't centralize them into a single server. To facilitate monitoring we're going to use different servers in Broker Mode which distributes the monitoring by its own method.
Distribution of remote testing model with agents in broker mode
In case of a primary hardware failure, we're required to setup an HA Server for security reasons. We're going to learn how to mount two servers: One ‘passive’, waiting in standby-mode for the active one to stop responding so it can start working. There are several ways to set them up in this way.
- The need to monitor a large volume of systems and manage them in a centralized way (more than 2500 agents). In order to do so, we're configuring different Pandora FMS Servers, coordinated by the system we call 'metaconsole'. They can be linearly scaled in this way.
The metaconsole model