JP2012212243A - Monitoring device and monitoring method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a monitoring device capable of appropriately detecting a state that a monitored device is down, and a monitoring method.SOLUTION: A monitoring device comprises: a noise removal part 225 that determines whether or not input image data input from a monitored device 100 is consistent with predetermined noise image data which means image data at the time of nonfailure, and that, when determining that the input image data is consistent with the noise image data, removes the noise image data from the input image data; a failure detection part 240 that determines whether or not the image data after removing the noise is consistent with failure image data which means predetermined image data of the monitored device at the time of failure, and that, when determining that the image data after removing the noise is consistent with the failure image data, determines that the monitored device is in a failure state; and a failure recovery procedure execution part 250 that, when the monitored device is determined to be in the failure state, executes predetermined failure recovery procedures with respect to the monitored device.

Description

  The present invention relates to a monitoring device and a monitoring method.

A monitoring device that monitors the down of a monitored device such as a server uses a polling process to perform an inquiry process of survival confirmation by ICMP (Internet Control Message Protocol) that is a protocol for confirming network communication. , Periodically for the monitored device. The monitoring device determines the survival status of the monitored device based on the response to the survival confirmation from the monitored device (polling method).
In addition, the monitored device proactively generates a trap by SNMP (Simple Network Management Protocol) for monitoring the network at predetermined time intervals with respect to the monitoring device. , Notify survival confirmation. Then, the monitoring device receives the survival confirmation from the monitored device to determine the survival status of the monitored device (survival notification method).

However, in such a polling method or alive notification method, it is necessary to shorten the alive check interval in order to increase the accuracy of detecting that the monitored device is down. As a trade-off for shortening the survival confirmation interval, there are disadvantages in that the load on the monitored device increases and the risk of false detection increases.
For this reason, some monitoring devices monitor a change in a signal continuously output by the monitored device, for example, a video output signal to a console screen, and detect a specific image pattern when the monitored device goes down. The image pattern unique when the monitored device goes down is, for example, the monitored device goes down and power is reset, then the screen is blacked out, then the monitored device vendor's logo is displayed, and then This is an image pattern in which a BIOS (Basic Input Output System) screen is displayed. Or, it is a display pattern such as a stack trace displayed at the time of OS (Operating System) panic, and a screen pattern such as a screen update stop. When the monitoring apparatus detects that the monitored apparatus is down using such an image pattern, the monitoring apparatus transmits a failure report such as a procedure for solving the failure to the administrator of the monitored apparatus (for example, Patent Document 1). reference).

JP 2006-65659 A

By the way, the image data of the screen saver of the monitored apparatus may have an image pattern similar to a specific pattern when it is down.
However, in the technique described in Patent Document 1, in the case of screen saver image data, the monitoring device may detect it as image data in a state where the monitored device is down. In such a case, there is a possibility that the monitoring apparatus may malfunction such as performing a repair process such as restarting the monitored apparatus that is operating normally.

  The present invention has been made in view of the above-described problems, and an object thereof is to provide a monitoring device and a monitoring method that can appropriately detect a down state of a monitored device.

  In order to achieve the above object, the monitoring apparatus according to the present invention determines whether or not the input image data input from the monitored apparatus matches noise image data that is predetermined non-failure image data. If it is determined that the input image data matches the noise image data, a noise removing unit that removes the noise image data from the input image data, and the noise image data output by the noise removing unit are removed. It is determined whether the noise-removed image data that is the image data matches failure image data that is predetermined image data at the time of failure of the monitored device, and the noise-removed image data Is determined to match the failure image data, a failure detection unit that determines that a failure has occurred in the monitored device, and the failure detection unit If serial is determined that failure to monitored devices has occurred, is characterized by comprising a fault recovery procedure execution unit for executing a disaster recovery procedure of the monitored device are predetermined, the.

  In the monitoring apparatus according to the present invention, the noise pattern is stored in association with the noise image data and information indicating an OS in which the noise image data is activated as information indicating the attribute of the noise image data. The noise removing unit determines whether the noise image data that matches the input image data is stored in the noise pattern database, and the noise image that matches the input image data. When the data is stored in the noise pattern database, the information indicating the OS activated in the monitored device as the information indicating the attribute stored in association with the noise image data; It is determined whether or not the attribute of the OS running on the monitoring device matches, and the stored If the attribute of the OS activated on the monitored device matches the attribute of the OS activated on the monitored device, it is determined that the input image data matches the noise image data, and the input When the image data matches the noise image data, the noise image data may be removed from the input image data.

  In the monitoring device according to the present invention, when the noise removal unit determines that the input image data is not the noise image data, the input image data changes for a predetermined first period or more. The noise image data may be removed from the input image data when it is determined that the input image data does not change for the preset period or longer.

  Further, in the monitoring device according to the present invention, the failure detection unit, when it is determined that the noise-removed image data output from the noise removal unit is the failure image data, the failure detection unit The polling process may be performed, and it may be determined that a failure has occurred in the monitored apparatus when there is no response to the polling process from the monitored apparatus.

  In the monitoring apparatus according to the present invention, the noise removal unit may determine whether the input image data is predetermined before determining whether the input image data matches the noise image data. It is determined whether or not the image data is black image continuously for a period of 2 or more, and the input image data is image data of black image continuously for a predetermined second period or more. If it is determined, it may be determined whether or not the input image data is the noise image data.

  In the monitoring device according to the present invention, the noise removing unit may acquire the input image data from the monitored device via a network.

  In the monitoring apparatus according to the present invention, when the failure detection unit determines that the noise-removed image data is the failure image data, the noise-removed image data is used as non-failure image data. As an alternative, a noise pattern learning unit may be provided which causes the noise pattern database to learn.

  In order to achieve the above object, according to the monitoring method of the present invention, the noise removing unit matches the input image data input from the monitored device with the noise image data that is predetermined non-failure image data. A noise removal step of removing the noise image data from the input image data, and a failure detection unit when the input image data is determined to match the noise image data. Whether or not the noise-removed image data that is the image data from which the noise image data is output matches the predetermined failure image data that is the image data at the time of failure of the monitored device A failure to determine that a failure has occurred in the monitored device when it is determined that the noise-removed image data matches the failure image data. A failure for executing a predetermined failure recovery procedure for the monitored device when the failure detecting unit determines that the failure has occurred in the monitored device by the detection step and the failure recovery procedure executing unit; A recovery procedure execution step.

  According to the present invention, the noise removal unit of the monitoring device removes non-failure image data from the image data output by the monitored device, and the failure detection unit detects non-failure from the image data output by the monitored device. The fault is detected based on the image data from which the image data is removed. When a failure is detected, the failure recovery procedure execution unit performs the failure recovery process on the monitored device, so that the down state of the monitored device can be detected appropriately.

1 is a schematic configuration diagram of a monitoring system 1 according to a first embodiment. It is a figure explaining an example of the data stored in noise pattern DB230 concerning the embodiment. It is a figure explaining an example of the data stored in panic pattern DB245 concerning the embodiment. 6 is a flowchart of noise pattern removal and panic processing according to the embodiment. It is a schematic block diagram of the monitoring system 1a which concerns on 2nd Embodiment. 6 is a flowchart of noise pattern removal and panic processing according to the embodiment. It is a schematic block diagram of the monitoring system 1b which concerns on 3rd Embodiment. It is a figure explaining an example of a structure of the physical layer 400 of the to-be-monitored apparatus 100 concerning the embodiment.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In addition, this invention is not limited to the embodiment which concerns, A various change is possible within the range of the technical thought.

[First Embodiment]
FIG. 1 is a schematic configuration diagram of a monitoring system 1 according to the present embodiment.
As shown in FIG. 1, the monitoring system 1 includes a monitored device 100 and a monitoring device 200. The monitored device 100 includes a communication unit 110, a video output unit 120, a keyboard signal input unit 130, a control unit 140, and a power supply unit 150. The monitoring apparatus 200 includes a video signal input unit 205, an image acquisition unit 210, a noise removal unit 225, a noise pattern DB 230, a failure detection unit 240, a panic pattern DB 245, a failure recovery procedure execution unit 250, a failure recovery procedure DB 255, a communication unit 260, A monitored device attribute DB 280 is provided.

First, the monitored apparatus 100 will be described. The monitored device 100 is a server, for example.
The communication unit 110 is a network card, a network connector, or the like, and has a LAN terminal, for example. The communication unit 110 is connected to the monitoring device 200 via the network 20 via the LAN cable 30. The communication unit 110 transmits information output from the control unit 140 to the monitoring device 200 via the network 20. The communication unit 110 receives the information transmitted from the monitoring device 200 via the network 20 and outputs the received information to the control unit 140.

  The video output unit 120 has, for example, a VGA output terminal. The video output unit 120 outputs a video signal output from the control unit 140.

  The keyboard signal input unit 130 includes a terminal to which an input signal from an input device such as a keyboard or a mouse (not shown) is input. The keyboard signal input unit 130 outputs a signal input from an input device such as a keyboard or a mouse to the control unit 140.

  The control unit 140 includes a CPU (Central Processing Unit), a RAM (Random Access Memory), a ROM (Read Only Memory), a HDD (Hard Disk Drive), and the like (not shown). The control unit 140 reads an OS (Operating System) held in an HDD (not shown), and outputs image data generated by the read OS to the video output unit 120. In addition, the control unit 140 performs processing of applications held in the OS and HDD based on the input signal output from the keyboard signal input unit 130.

  The power supply unit 150 supplies power input from the outside to each functional unit of the monitored device 100.

Next, the monitoring apparatus 200 will be described. The monitoring device 200 monitors the operation state of the monitored device 100. When the monitored device is inspected to be in the down state or the panic state, the monitoring device 200 notifies the administrator of the monitored device or the monitored device by e-mail or the like as will be described later.
The video signal input unit 205 receives a video signal output from the monitored device 100 and outputs the input video signal to the image acquisition unit 210. The video output unit 120 of the monitored apparatus 100 and the video signal input unit 205 of the monitoring apparatus 200 are connected by a video cable 10.

  The image acquisition unit 210 acquires the video signal output from the video signal input unit 205 and outputs input image data of the acquired video signal to the noise removal unit 225.

  In the noise pattern DB (database) 230, for example, image data imitating an image pattern at the time of down by various screen savers for each OS, and an image pattern displayed at the time of panic by the screen saver are stored in a still image format or a video format. Has been. Note that the image data simulating the image pattern at the time of down or panic by the screen saver is, for example, an OS operating on the monitored device 100 or an OS different from the OS operating on the monitored device 100 or This is image data of a screen saver on which an image simulating an image displayed at the time of panic is displayed. These screen savers may be installed by the administrator of the monitored apparatus 100 or may be provided by the OS as standard. Note that unique image data that resembles an image displayed when a screen saver is down or panicked is called a noise pattern. Note that the data stored in the noise pattern DB 230 may store image data corresponding to noise patterns obtained by connecting various types of monitored devices 100 in advance by an administrator of the monitoring device 200 or the like. .

FIG. 2 is a diagram for explaining an example of data stored in the noise pattern DB 230 according to the present embodiment. As shown in FIG. 2, the noise pattern DB 230 stores items of item number, image data, image data format, image data display duration, and OS type in association with each other.
The data format of the item number of the item is an integer, for example, a serial number. The item number can be arbitrarily stored in the noise pattern DB 230.
The image data of the item is unique image data (also referred to as false abnormal image or noise image data) resembling an image displayed when a screen saver is downed or panicked, for example, still image data or moving image data. . The data format of the image data is binary data. Storage of image data in the noise pattern DB 230 is essential.
The image data format of the item is, for example, bitmap, JPEG (Joint Photographic Experts Group), MPEG (Moving Picture Expert Group), or the like. The data format of the image data format is an integer or a character string. The image data format may be stored in advance by associating each format with an integer such that the bitmap is 1 and JPEG is 2, for example. Further, the image data format can be discriminated by, for example, an extension, and therefore is arbitrary for the noise pattern DB 230.
The image data display continuation time of the item is a period in which specific image data resembling an image displayed at the time of down or panic is displayed. The data format of the image data display duration is an integer. The unit of the image data display continuation time is, for example, second. Storage of the image data display duration in the noise pattern DB 230 is essential.
The item OS type is the type of OS in which the screen saver is used. The data format of the OS type is a character string, for example, an OS name including the OS version, such as OS1 and OS2. Storage in the OS type noise pattern DB 230 is essential.

  In the monitored device attribute DB (database) 280, the identification ID of the monitored device 100 and the attributes (including version) of the OS activated on the monitored device 100 are stored in association with each other.

The noise removal unit 225 determines whether or not the input image data output from the image acquisition unit 210 matches the image data stored in the noise pattern DB 230. When the input image data matches the image data stored in the noise pattern DB 230, the noise removing unit 225 further determines that the attribute of the image data stored in the noise pattern DB 230 is the monitored device attribute DB (database). It is determined whether or not the OS type activated by the monitoring apparatus 100 read from 280 matches the OS type.
The attribute of image data is the type of OS stored in the noise pattern DB 230 in association with the image data. That is, when the input image data is stored in the noise pattern DB 230 and the attribute matches the OS type of the monitored device 100, the noise removing unit 225 uses the image data output by the monitored device 100 as the screen saver image. Judge as data. If the attributes do not match, the noise removing unit 225 determines whether or not the detected image data is maintained for a predetermined period (first period). When the image data is switched within a predetermined period, it is determined that the input image data is screen saver image data. When it is determined that the image data is screen saver image data, the noise removal unit 225 removes the image data as a noise pattern. That is, the input image data is image data of a screen saver that resembles an image pattern when the monitored device 100 is in a down state or a panic state. For this reason, the noise removing unit 225 removes the image data as a noise pattern and outputs the noise-removed image data to the failure detecting unit 240.
On the other hand, when the input image data is not stored in the noise pattern DB 230, the noise removal unit 225 outputs the input image data to the failure detection unit 240 as it is.

The panic pattern DB (database) 245 stores an OS type, failure image data that is image data at the time of failure, a failure name, and the like in association with each other. The data stored in the panic pattern DB 245 may store image data corresponding to a failure acquired by connecting various types of monitored devices 100 in advance by an administrator of the monitoring device 200 or the like.
FIG. 3 is a diagram illustrating an example of data stored in the panic pattern DB 245 according to the present embodiment. As shown in FIG. 3, the panic pattern DB 245 stores an OS type, image data at the time of failure, and a failure name in association with each other. The image data is still image format or moving image format image data.

  The failure detection unit 240 receives the image data from which noise has been removed, which is output from the noise removal unit 225. The failure detection unit 240 determines whether or not the input noise-removed image data matches the image data stored in the panic pattern DB 245. When it is determined that the image data from which noise has been removed is stored in the panic pattern DB 245, the failure detection unit 240 determines that the monitored device 100 is in a down state or a panic state. The failure detection unit 240 reads information indicating the state of the monitored device 100 stored in association with the image data stored in the panic pattern DB 245, and sets the read information indicating the failure state of the monitored device 100 as a failure. The data is output to the recovery procedure execution unit 250. Note that the information indicating the state of the monitored device 100 stored in association with the image data may be, for example, the down state if the image data is image data when the image data is down, or the image data when the image data is panic. This is a failure state of the monitored apparatus 100 for each image data, such as a panic state.

  The failure recovery procedure execution unit 250 reads the failure recovery procedure from the failure recovery procedure DB 255 based on the information indicating the failure state of the monitored device 100 output from the failure detection unit 240. The failure recovery procedure execution unit 250 performs failure recovery processing on the monitored device 100 according to the read failure recovery procedure. The failure recovery procedure is a processing procedure such as polling the monitored apparatus 100 or notifying the administrator of the monitored apparatus 100 by e-mail, as will be described later.

  In the failure recovery procedure DB 255, a failure recovery procedure according to the failure status of the monitored device 100 is stored in advance for each OS and for each failure by the administrator of the monitoring device 200.

  The communication unit 260 is a network card, a network connector, or the like, and has a LAN terminal, for example. The communication unit 260 is connected to the monitored device 100 via the network 20. The communication unit 260 transmits information indicating the failure recovery process output from the failure recovery procedure execution unit 250 to the monitored apparatus 100 via the network 20. Further, the communication unit 260 receives information transmitted from the monitored device 100 via the network 20 and outputs the received information to the failure recovery procedure execution unit 250.

  Next, an example of noise pattern removal and panic processing will be described with reference to FIG. FIG. 4 is a flowchart of noise pattern removal and panic processing according to this embodiment.

(Step S1)
The image acquisition unit 210 of the monitoring device 200 receives image data output from the monitored device 100 via the video signal input unit 205. After step S1, the process proceeds to step S2.
(Step S2)
Next, the image acquisition unit 210 receives image data for one frame, and sequentially outputs the received input image data for one frame to the noise removal unit 225. After step S2, the process proceeds to step S3.

(Step S3)
Next, the noise removal unit 225 determines whether the input image data output from the pattern detection unit 215 matches the image data stored in the is pattern DB 230. That is, the noise removal unit 225 determines whether or not the monitored image 100 matches the specific image pattern that is output when the monitored device 100 is in a down state or a panic state. After step S3 ends, the process proceeds to step S4.

(Step S4)
When it is determined that the input image data matches the image data stored in the noise pattern DB 230 (step S4; Yes), the process proceeds to step S5. In this case, since it is determined that the input image data matches the image data stored in the noise pattern DB 230, the input image data is not the image data in the down state or the panic state, but the screen saver image data (noise pattern). there is a possibility.
If it is determined that the input image data does not match the image data stored in the noise pattern DB 230 (step S4; No), the process proceeds to step S8. In this case, since it is determined that the input image data does not match the image data stored in the noise pattern DB 230, the input image data may not be noise image data.

(Step S5)
Next, the noise removing unit 225 reads the OS attribute of the image data associated with the image data and stored in the noise pattern DB 230. In addition, the noise removing unit 225 reads the OS attribute of the monitoring device 100 that is associated with the identification ID of the monitored device 100 and stored in the monitored device attribute DB 280.
Next, the noise removing unit 225 determines again whether the OS attribute of the read image data matches the OS running on the monitored apparatus 100. For example, when the OS attribute of the image data detected by the noise removing unit 225 is read from the noise pattern DB 230, the attribute information of the corresponding image data is OS 2, and the monitored device 100 stored in the monitored device attribute DB 280. If the OS attribute is OS2, since the OS attributes match, it is determined that the screen saver image data by normal operation is operating on the input image data. For this reason, the noise removing unit 225 determines that the input image data is a noise pattern and removes the noise image data. After removing the noise pattern, the process returns to step S1.
On the other hand, if the OS attribute of the image data read from the noise pattern DB 230 does not match the OS running on the monitored device 100, the process proceeds to step S6. In this case, since the OS attribute of the image data read from the noise pattern DB 230 does not match the OS running on the monitored device 100, it is not noise image data (that is, an image indicating an actual server failure). Probability is high.

(Step S6)
When the input image data matches the image data stored in the noise pattern DB 230 and the OS attribute does not match, the noise removing unit 225 continues to check the image pattern for a predetermined period. The reason for this is to determine again whether or not the input image data is a screen saver that resembles the down state or panic state of the OS running on the monitored device 100. After step S6 ends, the process proceeds to step S7.

(Step S7)
In the case of a screen saver, the image data is repeated while changing periodically. For this reason, when image data changes, the noise removal part 225 judges that image data is image data of a screen saver (step S7; No), and returns to step S1.
On the other hand, if the image data does not change even after the predetermined period has elapsed (step S7; Yes), the monitored device 100 is likely to be in a down state or a panic state. The noise-removed image data is output to the failure detection unit 240, and the process proceeds to step S8. That is, the noise removal unit 225 removes noise pattern image data and image data that may be a noise pattern from the input image data output by the monitored apparatus 100, and the noise-removed image data after the noise pattern removal. Is output to the failure detection unit 240.

(Step S8)
Next, the failure detection unit 240 determines whether the noise-removed image data output from the noise removal unit 225 matches the image data stored in the panic pattern DB 245. When it is determined that the noise-removed image data matches the image data stored in the panic pattern DB 245, the failure detection unit 240 determines whether the monitored apparatus 100 is in a down state or a panic state based on the matched image data. Judge. That is, when the image data from which noise has been removed matches the image data in the down state, the failure detection unit 240 determines that the monitored device 100 is in the down state. Alternatively, when the image data from which noise is removed matches the image data in the panic state, the failure detection unit 240 determines that the monitored device 100 is in the panic state.
Next, the failure detection unit 240 reads information indicating the state of the monitored device 100 stored in association with the image data stored in the panic pattern DB 245, and indicates the read failure state of the monitored device 100. Information is output to the failure recovery procedure execution unit 250. The information indicating the failure state of the monitored device 100 is information indicating that the monitored device 100 is in a down state or information indicating that the monitored device 100 is in a panic state. After step S8 ends, the process proceeds to step S9.

(Step S9)
Next, the failure recovery procedure execution unit 250 reads the failure recovery procedure from the failure recovery procedure DB 255 based on the information indicating the failure state of the monitored device 100 output from the failure detection unit 240. For example, when the monitored device 100 output from the failure detection unit 240 is information indicating a down state, the failure recovery procedure execution unit 250 reads a failure recovery procedure in which the monitored device 100 is down from the failure recovery procedure DB 255. Alternatively, when the monitored device 100 output from the failure detection unit 240 is information indicating a panic state, the failure recovery procedure execution unit 250 reads a failure recovery procedure in which the monitored device 100 is in a panic state from the failure recovery procedure DB 255.
Next, the failure detection unit 240 performs failure recovery processing on the monitored device 100 according to the read failure recovery procedure.
The failure recovery procedure execution unit 250 communicates, with the communication unit 110 of the monitored apparatus 100, for example, an alive confirmation polling process by ICMP (Internet Control Message Protocol) according to the read failure recovery procedure. This is performed via the unit 260. The polling process is, for example, transmitting a ping signal or the like.
When there is a response from the monitored device 100 to the polling process for confirmation of survival, the failure recovery procedure execution unit 250 indicates that the video signal output unit 120 of the monitored device 100 may be broken, for example. Information is notified to the administrator of the monitored apparatus 100 by mail or the like via the communication unit 260. If there is no response from the monitored device 100 to the polling process for existence confirmation, the failure recovery procedure execution unit 250 determines that the monitored device 100 is in a down state or a panic state, and determines the determination result of the monitored device 100. The administrator is notified by e-mail or the like via the communication unit 260.

As described above, the noise removing unit 225 of the monitoring device 200 detects and removes a noise pattern by a screen saver that is similar to image data in a down state or a panic state from the acquired input image data. Then, the failure recovery procedure execution unit 250 determines whether a failure has occurred in the monitored apparatus 100 based on the image data from which noise has been removed, and executes the failure recovery procedure when a failure has occurred. Further, the failure recovery procedure execution unit 250 performs a polling process on the communication unit 110 of the monitored device 100 to determine whether the video signal output unit 120 of the monitored device 100 is in failure, in a down state or in a panic state. Judging.
As a result, even when a screen saver resembling image data at the time of down or panic is activated, the monitoring apparatus 200 can determine that the monitored apparatus 100 is operating normally, and erroneously detects the monitored apparatus 100. It is possible to prevent failure recovery processing. In addition, the monitoring device 200 can appropriately detect that the monitored device 100 is in a down state or a panic state, and can appropriately perform failure recovery processing when detected.

  In this embodiment, the example in which the noise removing unit 225 performs the attribute check between the OS activated in the monitored apparatus 100 and the OS type of the image data in step S5 has been described. Is not necessary. In this case, if Yes in step S4, the process proceeds to step S6 as it is, and whether or not the screen saver operation is performed may be determined by determining whether or not a predetermined period has elapsed (step S7).

  In addition, when the image data output from the monitored device 100 is, for example, a screen saver of image data resembling a panic state that includes a counter value that changes every moment, the moving image format image data that changes the counter value is converted into noise. It may be stored in the pattern DB 230 and the noise removing unit 225 may make a determination by comparing with the input image data. In the case of such a screen saver, for example, a fixed prefix character string is included before the counter value. For this reason, the noise removing unit 225 extracts a portion of the fixed prefix character string included before the counter value from the input image data, and determines whether or not the extracted image data is stored in the noise pattern DB 230. Then, noise may be removed.

  The identification ID of the monitored device 100 is acquired by, for example, acquiring the identification ID of the monitored device 100 superimposed on the input image data, for example, when the monitoring system 1 according to the present embodiment is constructed in advance. Based on the identification ID, the attribute of the OS activated on the monitored device 100 may be read from the monitored device attribute DB 280. In this case, the control unit 140 of the monitored device 100 superimposes the identification ID of the monitored device 100 on the video signal and outputs it from the video signal output unit 120. Alternatively, the monitoring device 200 may acquire the identification ID of the monitored device 100 via the communication unit 260, for example. The identification ID of the monitored device 100 may be, for example, a manufacturing number of the monitored device 100, a network IP (Internet Protocol) address, or a MAC (Media Access Control) address.

[Second Embodiment]
FIG. 5 is a configuration diagram of the monitoring system 1a according to the present embodiment.
As shown in FIG. 5, the monitoring system 1a includes a monitored device 100a, a monitoring device 200a, and a power supply control unit 300. The monitored device 100a includes a communication unit 110, a video output unit 120, a keyboard signal input unit 130a, a control unit 140, and a power supply unit 150. The monitoring device 200a includes a video signal input unit 205, an image acquisition unit 210a, a pattern detection unit 215, a timing unit 220, a noise removal unit 225, a noise pattern DB 230, a noise pattern learning unit 235, a failure detection unit 240a, a panic pattern DB 245, a failure. A recovery procedure execution unit 250a, a failure recovery procedure DB 255, a communication unit 260a, and a monitored device attribute DB 280 are provided. The same functional units as those in the first embodiment are denoted by the same reference numerals and description thereof is omitted.

First, the monitored device 100a will be described.
The keyboard signal input unit 130a includes a terminal to which an input signal from an input device such as a keyboard or a mouse (not shown) and an input signal from the monitoring device 200a are input. The keyboard signal input unit 130 a outputs a signal input from an input device such as a keyboard or a mouse and the monitoring device 200 a to the control unit 140.

  The power supply unit 150 supplies the power input from the power supply control storage 300 to each function unit of the monitored device 100.

  Next, the monitoring apparatus 200 will be described.

  The image acquisition unit 210 a acquires a video signal output from the video signal input unit 205 or the communication unit 260, and outputs input image data of the acquired video signal to the pattern detection unit 215.

  The timing unit 220 generates information indicating the time, and outputs the information indicating the generated time to the pattern detection unit 215.

The pattern detection unit 215 has a predetermined period (second period) based on information indicating the time output from the time measuring unit 220 by known image recognition from the input image data output from the image acquisition unit 210. A blackout state in which black signals are continuous is detected. The pattern detection unit 215 uses, for example, a pattern matching method as an image recognition method.
The reason for detecting a blackout state in which a black signal continues for a predetermined period is that, in the case of a general screen saver, after the display turns black once (this state is called blackout), the image data of the screen saver is Is output. Therefore, the pattern detection unit 215 detects this blackout in order to detect that the operation of the screen saver has started. When detecting the blackout, the pattern detection unit 215 sequentially outputs the input image data after the blackout detection to the noise removal unit 225 for each frame.

The failure detection unit 240a receives the noise-removed image data output from the noise removal unit 225. The failure detection unit 240 determines whether or not the input noise-removed image data matches the image data stored in the panic pattern DB 245. When it is determined that the noise-removed image data matches the image data stored in the panic pattern DB 245, the failure detection unit 240a determines that the monitored device 100 is in the down state or the panic state. The failure detection unit 240 reads information indicating the state of the monitored device 100 stored in association with the image data stored in the panic pattern DB 245, and sets the read information indicating the failure state of the monitored device 100 as a failure. The data is output to the recovery procedure execution unit 250.
Further, when it is determined that the image data from which noise has been removed does not match the image data stored in the panic pattern DB 245, the failure detection unit 240a is a new noise pattern image that is not stored in the noise pattern DB 230. The image data from which noise has been removed is output to the noise pattern learning unit 235.

  The noise pattern learning unit 235 learns the image data output from the failure detection unit 240a and causes the noise pattern DB 230 to learn.

  The failure recovery procedure execution unit 250a reads the failure recovery procedure from the failure recovery procedure DB 255 based on the information indicating the failure state of the monitored device 100 output from the failure detection unit 240. The failure recovery procedure execution unit 250a performs failure recovery processing on the power supply control device 300 and the monitored device 100 according to the read failure recovery procedure.

  The communication unit 260 a is connected to the monitored device 100 via the network 20. The communication unit 260 a transmits information indicating the failure recovery process output from the failure recovery procedure execution unit 250 to the monitored device 100 via the network 20. The communication unit 260a receives information transmitted from the monitored device 100 via the network 20, and outputs the received image data to the image acquisition unit 210a when the received information is image data.

  The power supply control device 300 is a device that supplies external power to the monitored device 100. The power supply control apparatus 300 controls the voltage supplied to the monitored apparatus 100 to be stopped or to be started by the control from the monitoring apparatus 200.

  Next, an example of noise pattern removal and panic processing will be described with reference to FIG. FIG. 6 is a flowchart of noise pattern removal and panic processing according to the present embodiment.

(Steps S101 and S102)
Steps S101 and S102 are performed in the same manner as steps S1 and S2 of the first embodiment. After step S102, the process proceeds to step S103.

(Step S103)
Next, the pattern detection unit 215 performs known image recognition based on the information indicating the time output from the time measuring unit 220 to determine whether or not the input image data output from the image acquisition unit 210a has been blacked out for a predetermined period or longer. To detect. The reason for this processing is that, in the case of a general screen saver, the screen saver is activated after the input image data is blacked out once. When blackout is detected, the pattern detection unit 215 outputs the input image data that has been input to the noise removal unit 225. After step S103 ends, the process proceeds to step S104.

(Steps S104 to S108)
Steps S104 to S108 are performed in the same manner as steps S3 to S7.

(Step S109)
Next, the failure detection unit 240a determines whether or not the noise-removed image data output from the noise removal unit 225 matches the image data stored in the panic pattern DB 245. After step S109 ends, the process proceeds to step S110.

(Step S110)
When it is determined that the input image data matches the image data stored in the panic pattern DB 245 (step S110; Yes), the process proceeds to step S112. In this case, the failure detection unit 240a determines that the monitored device 100a is in a down state or a panic state based on the matched image data.
If it is determined that the input image data does not match the image data stored in the panic pattern DB 245 (step S110; No), the process proceeds to step S111. In this case, since it is determined that the input image data does not match the image data stored in the panic pattern DB 245, the failure detection unit 240a determines that the monitored device 100a is in a normal operation state. Next, the failure detection unit 240a determines that the input image data is not yet stored in the noise pattern DB 230.

(Step S111)
When it is determined that the input image data does not match the image data stored in the noise pattern DB 230 (step S110; No), the failure detection unit 240a outputs the image data from which noise has been removed to the noise pattern learning unit 235. .
Next, the noise pattern learning unit 235 causes the noise pattern DB 230 to learn the noise-removed image data output from the failure detection unit 240a. After step S111 is completed, the process returns to step S101.

(Step S112)
When it is determined that the noise-removed image data matches the image data stored in the panic pattern DB 245 (step S110; Yes), the failure detection unit 240a causes the monitored device 100a to be down based on the matched image data. Judged to be in a state or panic state.
Next, the failure detection unit 240a reads information indicating the state of the monitored device 100a stored in association with the image data stored in the panic pattern DB 245, and indicates the read failure state of the monitored device 100a. The information is output to the failure recovery procedure execution unit 250a.

  Next, the failure recovery procedure execution unit 250a reads the failure recovery procedure from the failure recovery procedure DB 255 based on the information indicating the failure state of the monitored device 100a output from the failure detection unit 240a. The failure detection unit 240 performs failure recovery processing on the monitored device 100 according to the read failure recovery procedure. For example, the failure recovery procedure execution unit 250a generates a signal indicating a state where any key other than the software reset key combination is pressed or generates a signal indicating a state where the mouse is moved. Then, the failure recovery procedure execution unit 250a transmits a signal indicating that one of the generated keys has been pressed or a signal indicating the state of moving the mouse to the keyboard signal input unit 130a of the monitored device 100a. Whether or not the image data changes is confirmed. If the image data does not change after such processing, the failure recovery procedure execution unit 250a further performs the following processing.

The failure recovery procedure execution unit 250a transmits, for example, a software reset key combination signal to the keyboard signal input unit 130a of the monitored device 100a. Note that the software reset key combination signal is, for example, a signal when the Ctrl (control) key, the Alt key, and the Delete key are pressed simultaneously.
Next, after a predetermined time, the failure recovery procedure execution unit 250a performs, for example, a polling process on the communication unit 110 of the monitored apparatus 100a to check whether the software reset process has been executed. When there is no response from the monitored device 100a, the failure recovery procedure execution unit 250a transmits again, for example, a software reset key combination signal to the keyboard signal input unit 130a of the monitored device 100a. The predetermined time is a time equal to or longer than the time until the monitored device 100a performs software reset processing and restarts, for example, 120 seconds.
In addition, before transmitting the software reset key combination signal to the keyboard signal input unit 130 of the monitored device 100a, the failure recovery procedure execution unit 250a sends the monitored device 100 to the administrator of the monitored device 100a. Is transmitted via the network 20 by e-mail or the like. Then, after obtaining approval from the administrator, the failure recovery procedure execution unit 250a may transmit a software reset key combination signal to the keyboard signal input unit 130a of the monitored device 100a.

As described above, when the monitored device 100a cannot be restarted even if the software reset is performed on the keyboard signal input unit 130 of the monitored device 100a, or the failure recovery procedure executing unit 250a does not perform the software reset, Perform the following process.
The failure recovery procedure execution unit 250a transmits an instruction to stop the supply of power to the monitored device 100a to the power supply control device 300. As a result, power is not supplied from the power supply control device 300 to the monitored device 100a. Then, the monitored device 100a is forcibly shut down.
Next, the failure recovery procedure execution unit 250a transmits an instruction to start power supply to the monitored device 100a to the power supply control device 300 after a predetermined time. As a result, the monitored device 100a is restarted when power supply is resumed. Note that the predetermined time is a time when there is no problem even if the HDD is started after the rotation of the HDD of the monitored device 100a is stopped after the supply of power to the monitored device 100a is stopped. 120 seconds.
In addition, before transmitting an instruction to stop the supply of power to the monitored device 100a to the power supply control device 300, the failure recovery procedure execution unit 250a notifies the administrator of the monitored device 100a to the monitored device. Information indicating that the computer 100a is to be restarted is transmitted via the network 20 by e-mail or the like. Then, after obtaining approval from the administrator, the failure recovery procedure execution unit 250a may transmit an instruction to stop energization of the monitored device 100a to the power supply control device 300.

As described above, as a failure recovery procedure, first, the monitoring apparatus 200a moves the signal indicating that any key other than the software reset key combination is pressed on the monitored apparatus 100a or the mouse. A signal indicating the state is transmitted. Thereby, when the input image data from the monitored apparatus 100a is image data by the screen saver, the image data is changed by these processes. By this processing, the image data of the monitored device 100a is restored from the image data of the screen saver to the image data during the normal operation. As a result, the monitoring apparatus 200a can appropriately monitor the state of the monitored apparatus 100a.
If the input image data does not change even after the above processing is performed, the monitoring apparatus 200a transmits a software reset key combination signal to the monitored apparatus 100a and performs a software reset to the monitored apparatus 100a. I do. As a result, the monitoring device 200a can properly recover the failure of the monitored device 100a.
In addition, if the monitored device 100a does not restart even if a software reset is performed on the monitored device 100a, the monitoring device 200a stops supplying power to the monitored device 100a to the power supply control device 300. Send instructions to do. And after transmitting the instruction | indication which stops supply of electric power, after the predetermined time, the monitoring apparatus 200a restarted by restarting the power supply to the to-be-monitored apparatus 100a. As a result, the monitoring device 200a can properly recover the failure of the monitored device 100a.

  In the present embodiment, the example in which the noise pattern learning unit 235 causes the noise pattern DB 230 to learn the image data in step S111 has been described. However, the image data, the image data format, the image data display measurement time, the OS type, and the like are described. In association therewith, the noise pattern DB 230 may be made to learn.

  In the present embodiment, the example in which the monitoring device 200a acquires the image data output from the video signal output unit 120 of the monitored device 100a has been described. The image may be acquired from another output unit of the monitored apparatus 100a. For example, the image may be acquired via a LAN (Local Area Network) terminal of the communication unit 110. In this case, the control unit 140 of the monitored apparatus 100a transmits the image data to the network 20 via the communication unit 110. Then, the communication unit 260a of the monitoring device 200a receives the image data transmitted from the monitored device 100a via the network 20, and outputs the received image data to the image acquisition unit 210a. Alternatively, the control unit 140 of the monitored apparatus 100a transmits image data to the network 20 via a USB (Universal Serial Bus) terminal (not shown). Then, the communication unit 260a of the monitoring device 200a receives the image data transmitted from the monitored device 100a via the network 20, and outputs the received image data to the image acquisition unit 210a.

In this embodiment, the failure recovery procedure execution unit 250a first transmits a signal indicating that any key has been pressed to the monitored apparatus 100a or a signal indicating the state in which the mouse is moved. An example has been described in which it is determined again whether the monitoring device 100a is in a down state or a panic state. However, the re-determination of whether the monitored device 100a is in the down state or the panic state is not limited to this. For example, the failure recovery procedure execution unit 250a communicates with the monitored device 100a as in the first embodiment. If there is no response from the monitored device 100a even after performing the polling process to the unit 110, the failure recovery procedure execution unit 250a transmits, for example, a software reset key combination signal to the keyboard signal input unit 130a of the monitored device 100a. You may make it do.

[Third Embodiment]
Next, this embodiment will be described with reference to FIGS. 1 and 7. The same effect can be obtained by using the configuration of FIG.
In this embodiment, a plurality of virtual machines (Virtual Machines) are operating on the host OS of the monitored apparatus 100b. Note that a virtual machine is a virtual computer constructed of an actual computer on which the OS operates. Thus, a plurality of OSs can be executed in parallel by dividing one computer into virtual machines.

FIG. 7 is a schematic configuration diagram of the monitoring system 1b according to the present embodiment.
As shown in FIG. 7, the monitoring system 1b includes a monitored device 100b and a monitoring device 200b. The monitored device 100b includes a communication unit 110, video output units 120-1 to 120-n, a keyboard signal input unit 130, a control unit 140b, and a power supply unit 150. The monitoring device 200b includes a video signal input unit 205-1 to 205-n, an image acquisition unit 210b, a noise removal unit 225, a noise pattern DB 230, a failure detection unit 240, a panic pattern DB 245, a failure recovery procedure execution unit 250, and a failure recovery procedure. DB255, the communication part 260, and the to-be-monitored apparatus attribute DB280b are provided.

The control unit 140b of the monitored apparatus 100b outputs image data generated by the OS of the virtual machine to the video output units 120-1 to 120-n.
As described above, the reason for having a plurality of video signal output units 120-1 to 120-n is that some virtual machines operate using the entire screen of the host OS. In such a case, the image data output from the monitored device 100b is generally only the image data of the virtual machine VM ₁ displayed on a display device (not shown) connected to the monitored device 100b. The VM ₂ of the virtual machine operates in the background, and image data is not displayed on a display device (not shown). That is, when there is only one video output unit, only image data displayed on a display device (not shown) is output to the monitoring device 200b. For this reason, it is preferable that the monitored apparatus 100b includes a plurality of video output units 120-1 to 120-n.

  The video signal input units 205-1 to 205-n of the monitoring device 200b each receive the video signal output from the monitored device 100b, and output the input video signal to the image acquisition unit 210b. The video output unit 120-1 of the monitored apparatus 100b and the video signal input unit 205-1 of the monitoring apparatus 200 are connected by a video cable 10-1. Similarly, the video output unit 120-n of the monitored apparatus 100b and the video signal input unit 205-n of the monitoring apparatus 200 are connected by a video cable 10-n.

The image acquisition unit 210b acquires video signals output from the video signal input units 205-1 to 205-n, and outputs input image data of the acquired video signals to the noise removal unit 225.
In the monitored device attribute DB 280b, the identification ID of the monitored device 100 and the attributes (including the version) of the OS activated on the monitored device 100b are stored in association with each other. The monitored device attribute DB 280b stores an identification ID for each virtual machine activated in the monitored device 100b and an OS attribute for each virtual machine in association with each other.

FIG. 8 is a diagram illustrating an example of the configuration of the physical layer 400 of the monitored apparatus 100b according to the present embodiment. As illustrated in FIG. 8, the physical layer 400 of the monitored apparatus 100 b includes a lower layer 450 and an upper layer 410.
The lower layer 450 is a layer in which the host OS is operating.
The upper layer 410 is an application layer. In this embodiment, a plurality of virtual machines VM ₁ (420-1) to VM _n (420-n) are operating on the host OS. In such a virtual machine, the monitored apparatus 100b includes a plurality of video signal output units 120-1 to 120-n. For example, the monitored apparatus 100b outputs image data of the virtual machine VM ₁ (420-1) from the video signal output unit 120-1, and outputs image data of the virtual machine VM _n (420-n) to the video signal output unit 120. Output from -n. Alternatively, the monitored device 100b includes a plurality of communication units 110-1 to 110-n (not shown). The monitored device 100b outputs the image data of the virtual machine VM ₁ (420-1) from the communication unit 110-1 (not shown), and the image data of the virtual machine VM _n (420-n) is not shown. Output from.
When a plurality of virtual machines VM ₁ (420-1) to VM _n (420-n) are operating, each virtual machine VM ₁ (420-1) to VM _{n is} identified in order to identify which virtual machine the output is from. An identification ID is provided in the output of (420-n). For example, the identification ID to the virtual machine _VM 1 (420-1) is _{V 1,} the identification ID to the virtual machine _VM 2 (420-2) is _{V 2.}

For example, description will be made assuming that two virtual machines VM ₁ (420-1) and virtual machine VM ₂ (420-2) are operating on the managed device 100.
The video signal output unit 120 of the monitored apparatus 100 attaches an identification ID to the image data of the virtual machine VM ₁ (420-1) and the virtual machine VM ₂ (420-2) and outputs the image data to the monitoring apparatus 200.
The image acquisition unit 210 of the management apparatus 200 receives image data including identification IDs of the virtual machine VM ₁ (420-1) and the virtual machine VM ₂ (420-2) of the monitored apparatus 100.
The pattern detection unit 215, the noise removal unit 225, the failure detection unit 240, and the like of the monitoring device 200 monitor whether a failure has occurred in each virtual machine, as in the flowchart of FIG.

In step S5 of FIG. 4, the noise removal unit 225 determines whether the OS attribute of the image data read from the noise pattern DB 230 matches the OS attribute of the OS running on the virtual machine read from the monitored device attribute DB 280b. Determine again. When the OS attribute of the image data read from the noise pattern DB 230 matches the OS running on the virtual machine, the noise removal unit 225 determines that the input image data is a noise pattern and removes the noise image data. . After removing the noise pattern, the process returns to step S1.
On the other hand, if the OS attribute of the image data read from the noise pattern DB 230 does not match the OS running on the virtual machine, the process proceeds to step S6.
If the image data matches the image data stored in the panic pattern DB 245 in step S8 of FIG. 4, the failure detection unit 240 determines which virtual machine has a failure based on the identification ID attached to the image data. Determine if it has occurred. For example, when a failure has occurred in the virtual device VM ₁ , the failure detection unit 240 transmits an instruction to restart the virtual device VM ₁ to the monitored device 100 via the communication unit 260, for example. Alternatively, the failure detection unit 240 transmits information indicating that a failure has occurred in the virtual device VM ₁ to the administrator of the monitored device 100.

  In this embodiment, the example in which the virtual machine is identified using the identification ID attached to the image data has been described. However, the identification ID may not be used for identification. In this case, the image acquisition unit 210b of the monitoring apparatus 200b detects image data specific to the type and version of the virtual machine operating on the host OS during normal operation. Then, based on the detected result, the failure detection unit 240 may identify the virtual machine.

In the present embodiment, an example in which the monitored apparatus 100 includes a plurality of video signal output units has been described. However, the number of video signal output units may be one. In this case, the monitored apparatus 100 attaches identifiers to the image data of the virtual machines VM ₁ and VM ₂ and transmits them alternately, for example, in a time division manner.
Then, the video signal input unit 205b of the monitoring device 200b acquires the image data transmitted in this way, and detects whether a failure has occurred in each virtual machine using the acquired image data. Good.

  In the present embodiment, an example in which the monitored devices 100, 100a, and 100b are one has been described, but there may be a plurality of monitored devices. In this case, the monitoring devices 200 and 200a receive the image data output from the plurality of monitored devices in a time-sharing manner, or include a plurality of video signal input units like the monitoring device 200b shown in FIG. It may be.

In addition, you may make it implement | achieve a part of monitoring apparatus 200, 200a and 200b of FIG.1, FIG.5 and FIG.7 in embodiment mentioned above with a computer. In that case, the program for realizing the control function may be recorded on a computer-readable recording medium, and the program recorded on the recording medium may be read by a computer system and executed. Here, the “computer system” is a computer system built in the monitoring apparatus 200 and includes an OS and hardware such as peripheral devices. The “computer-readable recording medium” refers to a storage device such as a flexible medium, a magneto-optical disk, a portable medium such as a ROM and a CD-ROM, and a hard disk incorporated in a computer system. Furthermore, the “computer-readable recording medium” is a medium that dynamically holds a program for a short time, such as a communication line when transmitting a program via a network such as the Internet or a communication line such as a telephone line, In such a case, a volatile memory inside a computer system serving as a server or a client may be included and a program that holds a program for a certain period of time. The program may be a program for realizing a part of the functions described above, and may be a program capable of realizing the functions described above in combination with a program already recorded in a computer system.
Moreover, you may implement | achieve part or all of the monitoring apparatus 200a in the embodiment mentioned above, and 200b as integrated circuits, such as LSI (Large Scale Integration). Each functional block of the monitoring device 200 may be individually made into a processor, or a part or all of them may be integrated into a processor. Further, the method of circuit integration is not limited to LSI, and may be realized by a dedicated circuit or a general-purpose processor. Further, in the case where an integrated circuit technology that replaces LSI appears due to progress in semiconductor technology, an integrated circuit based on the technology may be used.

1, 1a, 1b ... monitoring system 1, 100 ... monitored device,
10 ... video cable, 20 ... network, 30 ... LAN cable,
110 ... communication unit, 120 ... video signal output unit,
130 ... Keyboard signal input unit, 140 ... Control unit, 150 ... Power supply unit,
200 ... monitoring device, 205 ... video signal input unit, 210 ... image acquisition unit,
215 ... Pattern detection unit, 220 ... Timekeeping unit, 225 ... Noise removal unit,
230 ... Noise pattern DB, 235 ... Noise pattern learning unit,
240 ... failure detection unit, 245 ... panic pattern DB,
250 ... Failure recovery procedure execution unit, 255 ... Failure recovery procedure DB, 260 ... Communication unit 280 ... Monitored device attribute DB

Claims (8)

It is determined whether or not the input image data input from the monitored device matches noise image data that is predetermined non-failure image data, and the input image data matches the noise image data If determined, a noise removing unit that removes the noise image data from the input image data;
The noise-removed image data that is the image data from which the noise image data output by the noise removing unit is removed matches the failure image data that is predetermined image data at the time of failure of the monitored device. A failure detection unit that determines that a failure has occurred in the monitored device when it is determined that the noise-removed image data matches the failure image data;
A failure recovery procedure execution unit that executes a predetermined failure recovery procedure of the monitored device when the failure detection unit determines that a failure has occurred in the monitored device;
A monitoring device comprising: A noise pattern database in which the noise image data is stored in association with information indicating an OS in which the noise image data is activated as information indicating the attribute of the noise image data;
The noise removing unit
Determining whether the noise image data matching the input image data is stored in the noise pattern database;
When the noise image data that matches the input image data is stored in the noise pattern database, it is activated by the monitored device as information indicating the attribute stored in association with the noise image data. Determining whether or not the information indicating the operating OS matches the attribute of the OS activated on the monitored device;
When the stored attribute of the OS activated on the monitored apparatus matches the attribute of the OS activated on the monitored apparatus, it is determined that the input image data matches the noise image data. The monitoring apparatus according to claim 1, wherein when the input image data matches the noise image data, the noise image data is removed from the input image data. The noise removing unit
When it is determined that the input image data is not the noise image data, it is determined whether or not the input image data is changed for a predetermined first period or more, and for a predetermined period or more, The monitoring apparatus according to claim 1, wherein when it is determined that the input image data does not change, the noise image data is removed from the input image data. The failure detection unit
When it is determined that the noise-removed image data output from the noise removing unit is the failure image data, a polling process is performed on the monitored apparatus, and a response from the monitored apparatus to the polling process is performed. The monitoring apparatus according to claim 1, wherein when there is no error, it is determined that a failure has occurred in the monitored apparatus. The noise removing unit
Before determining whether the input image data matches the noise image data,
It is determined whether or not the input image data is image data of a black image continuously for a predetermined second period or longer,
When it is determined that the input image data is continuously black image data for a predetermined second period or longer, it is determined whether or not the input image data is the noise image data. The monitoring apparatus according to claim 1, wherein the monitoring apparatus starts. The noise removing unit
The monitoring apparatus according to any one of claims 1 to 5, wherein the input image data is acquired from the monitored apparatus via a network. Noise that causes the noise pattern database to learn the noise-removed image data as non-failure image data when the failure detection unit determines that the noise-removed image data is the failure image data. The monitoring apparatus according to claim 2, further comprising a pattern learning unit. The noise removing unit determines whether or not the input image data input from the monitored apparatus matches noise image data that is predetermined non-failure image data, and the input image data is the noise If it is determined that the image data matches, a noise removal step of removing the noise image data from the input image data;
The failure in which the failure detection unit is the image data from which the noise image data output from the noise removal unit is removed and the noise-removed image data is predetermined image data at the time of failure of the monitored device Determining whether or not the image data matches, and when determining that the noise-removed image data matches the failure image data, a failure detection step of determining that a failure has occurred in the monitored device;
A failure recovery procedure execution step for executing a predetermined failure recovery procedure for the monitored device when the failure detection unit determines that the failure has occurred in the monitored device. When,
The monitoring method characterized by including.

Images