JP4909830B2 - Server application monitoring system and monitoring method - Google Patents

Description

  The present invention relates to a server application monitoring system and a monitoring method, and more particularly to a server application monitoring system and a monitoring method for monitoring the performance of a server application in a client / server type system.

  As a method for monitoring the performance of a server application in a client / server type system, a measurement agent device installed with a program called an agent or a probe that simulates a processing request or transaction for the server application is installed on the client side where the user of the server application is located. There is known a server application monitoring system in which a server monitoring apparatus deployed at a base and installed with a server monitoring manager program controls the measurement agent apparatus. Such a server application monitoring system evaluates the performance of a server application by measuring the response time of a server application to a processing request from a measurement agent device or a transaction, and the time required for a series of processing for the request. This method can evaluate the performance of the server application as the actual user experience speed.

  The above-described conventional technology has an advantage that when a problem occurs in the server application, a plurality of measurement agent programs periodically measure the performance of the server application, so that the occurrence can be detected immediately. Have.

Further, as another conventional technique for monitoring the performance of the server application, for example, by using the technique using the measurement agent program described above in cooperation with the network monitoring system described in Patent Document 1, or A technique is known in which a problem occurrence location is specified by using in cooperation with a network device operation monitoring system as described in Patent Document 2. These prior arts can support the system administrator so that the problem can be solved at an early stage by notifying the system administrator of the identified problem.
Special table 2003-508849 gazette JP 2004-104540 A

  The prior art using the measurement agent program described above is a method in which the measurement agent device only periodically measures, and the server monitoring device cannot immediately determine the cause of the problem. However, there is a problem that the administrator of the monitored system must perform all the work to identify the cause of the problem and to recover from the problem. is doing.

  In addition, the conventional technology linked with the network monitoring system and the network equipment operation monitoring system can determine the cause of the problem, but the monitoring result used to determine the cause of the problem is the result of the user's experience. There is a problem that there is a case where the speed does not coincide with the result of evaluating the performance of the server application. In addition, these conventional technologies have a problem that, after a problem occurs, measurement from the measurement agent device to the server application may place an additional load on the network, server, or server application in which the problem occurs. Has a problem.

  The object of the present invention is to solve the above-mentioned problems of the prior art, monitor the performance of server applications without using other monitoring systems such as a network monitoring system and a network equipment operation monitoring system, and It is an object of the present invention to provide a server application monitoring system and a monitoring method capable of determining the cause of the problem when detected and recovering from the problem.

According to the present invention, the object is a server application monitoring method in a server application monitoring system comprising a plurality of servers that execute a plurality of server applications, a plurality of measurement agent devices, and a server monitoring device that monitors the servers. ,
The server monitoring device
For the combination of the server application and the measurement agent device, the priority, threshold value and measurement interval are stored in advance,
According to the measurement interval, repeatedly transmits a measurement instruction of the specified server application to a plurality of the measurement agent devices,
The measurement agent device includes:
In response to a measurement instruction from the server monitoring device, measure a response time from requesting processing to the instructed server application until receiving a result of the requested processing, and measuring the measured response time, Sent to the server monitoring device,
The server monitoring device
Repeatedly receiving the response time from a plurality of the measurement agent devices,
By comparing the response time received from the measurement agent device with the threshold value, the measurement agent device that has transmitted the response time exceeding the threshold value, and the server application as a measurement target, Extract the first combination containing
Obtaining one or more second combinations related to the first combination, including the server application and the measurement agent device;
Calculating an average value of the priority of the first combination and the priority of one or more of the second combinations;
By comparing the calculated average value with the priority of the first combination and the priority of one or more of the second combinations, a third combination having a higher priority than the average value is obtained. Determining from a first combination and one or more of the second combinations;
The response time measurement instruction for the third server application included in the third combination is sent to the third combination included in the third combination at a time interval shorter than the application measurement interval corresponding to the third combination. Sent to the measurement agent device,
Determining a fourth combination having a lower priority than the average value from the first combination and the one or more second combinations;
This is achieved by suspending the repeated transmission of the response time measurement instruction for the fourth server application included in the fourth combination to the fourth measurement agent device included in the fourth combination .

  According to the present invention, the occurrence of a problem related to the performance of a server application is detected without using another monitoring system such as a network monitoring system or a network equipment operation monitoring system, and the occurrence of the problem is detected. Immediately determine the cause of the problem, eliminate the need for the system administrator to perform recovery work, and perform recovery from the problem. Even if a problem occurs, no additional load is placed on the network, server, server application, etc. In addition, monitoring of the server application can be continued, and the burden on the operation and monitoring of the server application can be reduced.

  DESCRIPTION OF EMBODIMENTS Hereinafter, embodiments of a server application monitoring system and a monitoring method according to the present invention will be described in detail with reference to the drawings.

  FIG. 1 is a block diagram showing the overall configuration of a server application monitoring system according to an embodiment of the present invention.

  The server application monitoring system according to the embodiment of the present invention is configured by connecting a server monitoring apparatus 100, a measurement agent apparatus 120, a server resource management apparatus 140, one or more servers 150, and one or more clients 160 to a network 130. ing.

  The server application monitoring system shown in FIG. 1 is shown as having one measurement agent device 120, but a plurality of measurement agent devices 120 may be provided, and FIG. For the sake of simplicity, only one measurement agent device 120 is shown. For simplification of the drawing, FIG. 1 shows only one combination of two servers 150 and one server resource management device 140, but three or more servers 150 and one server resource management device 140 are shown. It is possible to provide a combination with the server resource management apparatus or to provide a plurality of such combinations.

  In the server application monitoring system configured as described above, the server monitoring apparatus 100 that operates as a monitoring manager is a server application that operates on the server 150 that is the monitoring target in the measurement agent apparatus 120 that operates as a monitoring agent according to a predetermined monitoring schedule. 151, the measurement agent device 120 performs communication determined with the server application 151, measures the time required for the communication, and measures the time required for the communication measured as a result of the monitoring by the server monitoring device 100. To report. Then, the server monitoring apparatus 100 detects the occurrence of the problem by comparing the reported monitoring result with a predetermined threshold value, and changes the monitoring schedule to perform the monitoring in order to identify the problem location The scope of the problem is identified by analyzing the results of the monitoring performed by changing the monitoring schedule, the monitoring schedule is adjusted according to the identified scope of the problem, and the problem is identified according to the identified scope of the problem. The server resource management apparatus 140 is instructed to execute the process for recovery.

  The server monitoring apparatus 100 includes a processor (CPU) 104 and a memory 101, and is connected to a network 130 via a communication interface 102. The server monitoring apparatus 100 is connected to a storage device 109 such as a hard disk drive via a disk interface 103. The processor 104 operates the server monitoring apparatus 100 as a monitoring manager by executing the server monitoring manager program 105 stored in the storage device 109 and loaded into the memory 101 and stored in the memory 101.

  The server monitoring manager program 105 includes a measurement control unit 106 that executes measurement control processing, a measurement result analysis unit 107 that executes measurement result analysis processing, and a recovery unit execution unit 108 that executes recovery means execution processing. The server monitoring manager program 105 reads and writes the measurement information table group 110 and the measurement result table 111 stored in the storage device 109 when executing each process.

  The measurement agent device 120 includes a processor 122 and a memory 121, and is connected to the network 130 via the communication interface 124. The measurement agent device 120 is connected to a storage device 127 such as a hard disk drive via a disk interface 123. The processor 122 operates the measurement agent device 120 as a monitoring agent by executing the server monitoring agent program 125 stored in the storage device 127 and loaded into the memory 121 and stored in the memory 121.

  The server monitoring agent program 125 includes a measurement execution processing unit 126 that executes measurement execution processing. When the measurement execution processing is executed, the server monitoring agent program 125 reads and writes the measurement data table 128 stored in the storage device 127.

  Although the detailed configuration of the client 160 is not shown, the client 160 is an information processing device represented by a PC or the like, and includes a CPU, a memory, a communication interface, a storage device such as a hard disk, an output device such as a display device, a keyboard, a mouse, and the like. The input device is configured. A client application is loaded in the memory of the client 160, and the CPU controls the execution of the client application 161 and provides a service using the server application 151 in the server 150 to the user.

  FIG. 2 is a block diagram illustrating a configuration example of the server 150. The server 150 includes a processor 201 and a memory 200 and is connected to the network 130 via the communication interface 203. The server 150 is connected to a storage device 204 such as a hard disk drive via the disk interface 202. The processor 201 is stored in the storage device 204 and executes a server application program 205, a Web server program, a Web application program 207 that operates on the Web server program, and the like loaded into the memory 200 and stored in the memory 200. .

  The server application monitoring system according to the embodiment of the present invention targets these application program 205, Web application program 207, and the like.

  FIG. 3 is a block diagram illustrating a configuration example of the server resource management apparatus 140. The server resource management device 140 illustrated in FIG. 3 is configured by the load distribution device 300.

  The load distribution apparatus 300 includes a load distribution processing unit 301 that performs communication load distribution for the server 150 that is received from the network 130, a server monitoring unit 302 that monitors the operation status of the server 150, and a server resource control unit that controls resources allocated to the server 150. 303, a server resource control means providing unit 304 that provides server resource control means such as a command line interface to the server administrator and external software.

  The example of the load balancer 300 shown in FIG. 3 is shown as being realized by hardware, but both load balancers realized by hardware and load balancers realized by software are widely known. Exist. When the load balancer 300 is configured by software, the load balancer 300 includes a processor and a memory, is connected to the network 130 via a communication interface, and is connected to a hard disk drive or the like via a disk interface. Each functional unit described above is constructed by connecting to a storage device and executing a program stored in the storage device and loaded into the memory and stored in the memory. That's fine.

  In the above description, the load distribution device 300 has been described as an example of the server resource management device 140. However, the server resource management device 140 is provided by a logical partition function provided by a server, a virtual server function provided by software, or a network device. It is possible to use software and hardware for managing server resources that exist widely and generally, such as a bandwidth control function. The server resource management apparatus 140 constituting the embodiment of the present invention is premised only on providing server resource control means to external software, and its configuration and implementation method are not limited.

  FIG. 4 is a diagram showing the tables constituting the measurement information table group 110 stored in the storage device 109 of the server monitoring apparatus 100.

  The measurement information table group 110 includes a measurement target table 401 that defines information of a server application 151 that is a target for measuring communication time, a measurement setting table 402 that defines setting information for each measurement, A measurement interval table 403 for storing a schedule, a problem management table 404 for storing problem management information when the occurrence of a problem is detected, and a recovery means table 405 defining information on means for recovering from the problem Consists of.

  FIG. 5 is a diagram showing a configuration example of the measurement target table 401 shown in FIG.

  The measurement target table 401 is a table defining information of the server application 151 as a monitoring target for measuring the communication time, and includes a target application column 504 for registering the name or identifier of the server application 151 to be measured. The target server column 503 for registering the IP address or host name of the server 150 on which the server application 151 is executed, and the target network for registering the network address or network name of the server network to which the server 150 on which the server application 151 is executed belongs A column 502 and a target ID column 501 for registering a target ID composed of numbers or symbols that uniquely identify a set of definitions. The definition of columns other than the target ID column 501 is registered by the user of the server monitoring manager program 105.

  For example, in order to register the server application 151 having the name AP1 operating on the server 150 having the host name Server1 in the network having the network name Network1 as the measurement target, Network1 is registered in the target network column 502, and Server1 is registered in the target server column 503. AP1 is defined in the target application column 504. In the target ID column 501, T1 is assigned as a target ID by the server monitoring manager program 105.

  FIG. 6 is a diagram showing a configuration example of the measurement setting table 402 shown in FIG.

  The measurement setting table 402 is a table that defines setting information for each measurement determined by a combination of the measurement agent device 120 that performs measurement and the server application 151 that is the measurement target, and is the measurement target illustrated in FIG. A target ID column 602 for registering the target ID of the target ID column 501 defined in the table 401, a measurement execution agent column 603 for registering the IP address or host name of the measurement agent device 120 that performs measurement, and how much the measurement is performed A priority column 604 for registering priority expressed as an integer from the highest 10 to the lowest 1 for evaluating whether to perform priority, and a threshold time for evaluating the communication time obtained as a result of the measurement A threshold string 605 to be registered, a measurement interval string 606 for registering an interval for performing the measurement, and one set Constituted by the measurement ID column 601 for registering a measurement ID consisting of uniquely identifying number or symbol definitions. The definition of columns other than the measurement ID column 601 is registered by the user of the server monitoring manager program 105.

  For example, a measurement target assigned T1 as a target ID is measured by the measurement agent device 120 having the host name Agent1, its priority is 10, its threshold is 0.3 seconds, and the measurement execution interval is 300. To register as seconds, T1 is defined in the target ID column 602, Agent1 is defined in the measurement execution agent column 603, 10 is defined in the priority column 604, 0.3 seconds is defined in the threshold column 605, and 300 seconds is defined in the measurement interval column 606. . In the measurement ID column 601, M1 is assigned as a measurement ID by the server monitoring manager program 105.

  FIG. 7 is a diagram showing a configuration example of the measurement interval table 403 shown in FIG.

  The measurement interval table 403 is a table that the server monitoring manager program 105 creates, sets, and changes in order to determine the interval at which the measurement control unit 106 of the server monitoring manager program 105 instructs the execution of measurement. The measurement interval table 403 includes a measurement ID column 701 for setting the measurement ID of the measurement ID column 601 shown in FIG. 6 and the measurement execution for each measurement defined in the measurement setting table 402 shown in FIG. A measurement execution on / off column 702 for setting presence / absence by a character string On or Off, a measurement interval column 703 for setting an interval for performing measurement, and a next measurement time column 704 for setting a time for executing the next measurement Composed.

  For example, when measurement with M1 assigned as the measurement ID is executed, the interval for executing the measurement is 300 seconds, and the time for executing the next measurement is 12:00, M1 is measured in the measurement ID column 701. On is set in the on / off column 702, 300 seconds is set in the measurement interval column 703, and 12:00 is set in the next measurement time column 704.

  FIG. 8 is a flowchart for explaining the operation of the measurement control process in the measurement control unit 106 of the server monitoring manager program 105 of the server monitoring apparatus 100 shown in FIG. 1. Next, referring to the flowchart shown in FIG. A processing operation in the measurement control unit 106 of the monitoring manager program 105 will be described.

(1) When the processing is started, the measurement control unit 106 refers to the time set in the next measurement time column 704 of the measurement interval table 403 described with reference to FIG. 7, and until the time closest to the current time is reached. stand by. Then, when that time comes, or when that time has already passed, the process proceeds to the next process without waiting. For example, if the current time is 11:59, the process waits until 12:00 when the measurement ID M1 is measured in the measurement interval table 403 (step 801).

(2) Next, referring to the measurement interval table 403 shown in FIG. 7, the measurement setting table 402 shown in FIG. 6, and the measurement target table 401 shown in FIG. 5, the host of the measurement agent device 120 that performs measurement The name or IP address, the host name or IP address of the server 150 to be measured, and the name or identifier of the server application 151 to be measured are determined. For example, when determining the measurement with the measurement ID M1 in the measurement interval table 403, the host name Agent1 of the measurement agent device 120 is determined from the measurement setting table 402. Since the target ID is T1, the host name Server1 of the measurement target server 150 and the server application name AP1 of the measurement target are determined from the measurement target table 401 (step 802).

(3) The measurement control unit 106 instructs the measurement agent device 120 that executes the measurement determined in step 802 to execute the measurement. For the measurement instruction, a communication packet including the host name or IP address of the measurement target server 150 determined in the process of step 802 and the name or identifier of the measurement target server application 151 is transmitted to the measurement agent device 120. To implement. For example, when the measurement interval table 403 instructs the execution of the measurement with the measurement ID M1, the communication packet including the host name Server1 of the server 150 and the name AP1 of the server application 151 to the measurement agent device 120 having the host name Agent1. Is transmitted (step 803).

(4) Next, the next measurement time column 704 in the measurement interval table 403 is updated. That is, the time obtained by adding the current time to the measurement interval corresponding to the measurement instructed to be executed in step 803 is set as the next measurement time. For example, if the measurement ID of the measurement interval table 403 is instructed to execute measurement with M1, 12:05 is set as the next measurement time of the measurement (step 804).

(5) Thereafter, the process waits until a communication packet including the measurement result arrives from the measurement agent device 120 that has instructed execution of the measurement in the process of step 803. Then, the communication packet that has arrived is received, and the measurement result included in the packet is acquired (step 805).

(6) The measurement result analysis unit 107 of the server monitoring manager program 105 is notified of the measurement ID for which the measurement has been performed and the measurement result acquired in step 805, and then acquired in the measurement result table 110 through step 805. The measured result is stored, and the process here is terminated (step 806).

  FIG. 9 is a diagram illustrating a configuration of the measurement result table 111 stored in the storage device 109 of the server monitoring apparatus 100.

  The measurement result table 111 includes a measurement ID column 901 that indicates the measurement ID of the measurement that has been performed, a measurement time column 902 that stores the time at which the measurement was performed, and a measurement result column 903 that stores the measurement result. . Each time a new measurement is executed, a measurement time column 902 and a measurement result column 903 are added to the measurement result table 111, and the measurement time and the measurement result are stored in the added column.

  FIG. 10 is a flowchart for explaining the operation of the measurement execution process in the measurement execution unit 126 of the server monitoring agent program 125 in the measurement agent device 120 shown in FIG. 1, which will be described next. The measurement execution processing here is performed when the server monitoring agent program 125 is instructed to execute measurement by the server monitoring manager program 105 in step 805 of the flow described with reference to FIG. 8, that is, the host name of the measurement target server 150. Alternatively, it is executed when a communication packet including the IP address and the name or identifier of the server application 151 to be measured is received.

(1) The measurement execution unit 126 acquires and acquires the host name or IP address of the measurement target server 150 and the name or identifier of the measurement target server application 151 from the communication packet received from the server monitoring agent program 125. Using the measurement target information, a measurement method for the measurement target server application 151 is determined with reference to a measurement data table 128 described later (step 1001).

(2) Thereafter, the measurement execution unit 126 acquires and holds the current time as the measurement start time t1, and communicates with the server application 151 of the measurement target server 150 according to the method determined in step 1001. Implement (step 1002).

(3) Wait for a response from the server application 151 to the communication performed in the process of step 1002, receive a response from the server application 151 to the communication, and then acquire and hold the current time as the measurement end time t2 ( Step 1003).

(4) Calculate the time obtained by subtracting the measurement start time t1 acquired in the process of step 1002 from the measurement end time t2 acquired in the process of step 1003, and set the communication packet including the measurement result that is the time as the server monitoring manager. It transmits to the program 105, and a process here is complete | finished (step 1004).

  FIG. 11 is a diagram illustrating a configuration example of the measurement data table 128 stored in the storage device 127 of the measurement agent device 120.

  The measurement data table 128 is a table that defines information on a method of performing measurement for the server application 151 that is to measure the communication time, and is a target for registering the name or identifier of the server application 151 to be measured. The application column 1102 includes a target server column 1101 for registering the IP address or host name of the server 150 on which the server application 151 is executed, and a measurement method column 1103 for registering a measurement method for the server application 151.

  The measurement method registered in the measurement method column 1103 includes a communication packet transmitted from the server monitoring agent program 125 to the server application 151, and a communication packet transmitted from the server monitoring agent program 125 to the server application 151 when executed. Is a command or program that can send.

  FIG. 12 is a diagram showing a configuration example of the problem management table 404 shown in FIG.

  The problem management table 404 is a table for storing measurement information determined to have a problem by the measurement result analysis unit 107 of the server monitoring manager program 105 of the server monitoring apparatus 100. A problem measurement ID column 1201 to be stored, a state column 1202 to store the internal state of the measurement result analysis unit 107 for the problem, and a relationship to store a measurement ID of a measurement related to the measurement determined to have a problem A measurement ID column 1203, a relationship type column 1204 for storing information indicating the relationship between a measurement determined to have a problem and a measurement related to the measurement, and a measurement problem related to the measurement ID column 1203 are employed. The measurement result column 1205 stores the result of comparing the latest measurement result and the threshold value of the measurement setting table 402.

  Here, the related measurement will be described in detail. Assume that a plurality of measurement agent devices 120 execute measurement for one server application 151. At this time, there are a plurality of measurements that are combinations of the server application 151 and the measurement agent device 120. In this case, there is a relationship that the server applications 151 to be measured are the same.

  In addition, when a plurality of server applications 151 are executed on one server 150, a plurality of measurement agent devices 120 execute measurement for the plurality of server applications 151. At this time, there are a plurality of measurements that are combinations of the server application 151 and the measurement agent device 120. In this case, there is a relationship that the servers 150 to be measured are the same. That is, it is related in that the server 150 that operates among the attributes of the server application 151 is the same.

  Similarly, when a plurality of servers 150 exist in one server network and a plurality of server applications 151 are executed on the servers 150, a plurality of measurement agent devices 120 are associated with the plurality of server applications 151. Shall perform the measurement. At this time, there are a plurality of measurements that are combinations of the server application 151 and the measurement agent device 120. In this case, there is a relationship that the server networks to be measured are equal. That is, it is related that the server networks to which the server application 151 belongs are equal.

  The server application monitoring system according to the described embodiment of the present invention handles the three relations as described above. In the example illustrated in FIG. 12, the related measurement as described above is illustrated as extracted from the measurement target table 401 illustrated in FIG. 5 and the measurement setting table 402 illustrated in FIG. 6.

  The status stored in the status column 1202 of the problem management table 404 shown in FIG. 12 is set to one of the values “problem being confirmed” or “recovery being checked”, and the relationship stored in the relationship type column 1204 is displayed. The information to be expressed is either “the target network column 502 of the measurement target table 401 is equal”, “the target server column 503 of the measurement target table 401 is equal”, or “the target application column 504 of the measurement target table 401 is equal”. Value is set. In addition, as a result of comparison with the threshold value stored in the measurement result column 1205, one of the values “exceeding the threshold value”, “below the threshold value”, and “not measured” is set. The

  FIG. 13 is a flowchart for explaining the operation of the measurement result analysis process in the measurement result analysis unit 107 of the server monitoring manager program 105 in the server monitoring apparatus 100 shown in FIG. 1, and this will be described next. This measurement result analysis process is executed when the measurement control unit 106 is notified of the measurement ID and the measurement result.

(1) When the processing here is started, the measurement result analysis unit 107 searches the related measurement ID column 1203 of the problem management table 404 illustrated in FIG. 12 and the measurement ID notified from the measurement control unit 106 Is registered (step 1301).

(2) If the target measurement ID is not registered in the problem management table 404 in the determination in step 1301, the measurement result notified from the measurement control unit 106 is the threshold value of the measurement setting table 402 shown in FIG. It is determined whether or not the threshold value in the column 605 has been exceeded. If the measurement result is equal to or less than the threshold value, the measurement result analysis process is terminated (step 1302).

(3) If it is determined in step 1302 that the measurement result exceeds the threshold value, it is determined that a new problem has occurred. Therefore, a new problem registration process to be described later is executed according to the flow shown in FIG. (Step 1303).

(4) When the target measurement ID is registered in the problem management table 404 in the determination in step 1301, it is determined that the measurement result related to the problem already registered is received by the process in step 1301. Therefore, the problem handling process described later is executed according to the flow shown in FIG. 15 (step 1304).

  FIG. 14 is a flowchart for explaining the operation of the new problem registration process in step 1303 of the flow shown in FIG. This new problem registration process is a process executed by the measurement result analysis unit 107 to register the problem in the problem management table 404 when it is determined that a new problem has occurred. By this problem new registration process, the related measurement is immediately executed in order to identify the problem location of the problem where the new problem has occurred.

(1) When this processing is started, the measurement result analysis unit 107 sets the measurement ID of the target measurement in the problem measurement ID column 1201 of the problem management table 404 illustrated in FIG. “Problem confirmation in progress” is set (step 1401).

(2) Next, it is set to repeat the processing of Step 1403 and Step 1404 for the measurement related to the target measurement (Steps 1402 and 1405).

(3) The measurement ID of the measurement related to the related measurement ID column 1203 of the problem management table 404 is set, the information indicating the relationship is set in the related type column 1204, and “measurement not performed” is set in the measurement result column 1205 (Step 1403).

(4) In order to immediately perform the related measurement, the current time is set in the next measurement time column 704 corresponding to the measurement ID of the related measurement in the measurement interval table 403 shown in FIG. 7 (step 1404).

  FIG. 15 is a flowchart for explaining the operation of the problem handling process in step 1304 of the flow shown in FIG. This problem handling process is a process executed by the measurement result analysis unit 107 in order to identify a problem location of the problem or confirm that the problem has been recovered when a problem has already occurred. .

(1) When this processing is started, the measurement result analysis unit 107 refers to the status column 1202 of the problem management table 404, and the internal state of the measurement result analysis unit 107 for the problem is “under problem confirmation”. Or “recovery being confirmed” (step 1501).

(2) If the internal state is “under problem confirmation” in the determination in step 1501, the measurement result of the target measurement is compared with the threshold value in the threshold value column 605 of the measurement setting table 402; When the comparison result exceeds the threshold value, “exceeding threshold value” is set in the measurement result column 1205 of the problem management table 404, and when the comparison result is less than the threshold value, “ The threshold management "is set and the problem management table 404 is updated (step 1502).

(3) Next, referring to the measurement result column 1205 of the problem management table 404, it is determined whether or not there is a related measurement that is “not measured”, and the relationship that is “not measured” is determined. If there is a measurement to be performed, the related processing cannot be performed here until the measurement is completed, and thus the problem handling processing is terminated (step 1503).

(4) In the determination in step 1503, when there is no related measurement that is “not measured”, it is determined that the measurement for analyzing the problem range is completed by the processing in step 1501 and step 1503. Therefore, the problem analysis process described later with reference to FIG. 16 is executed (step 1504).

(5) Thereafter, in order to adjust the measurement execution interval according to the analysis result of the problem analysis process in step 1504, a measurement adjustment process described later with reference to FIG. 17 is executed (step 1505).

(6) If it is determined in step 1501 that the internal state is “confirming recovery”, recovery confirmation processing described later with reference to FIG. 18 is executed (step 1506).

(7) Next, similarly to the determination in step 1503, the measurement result column 1205 of the problem management table 404 is referred to, and it is determined whether there is any related measurement that is “not measured”. If there is a related measurement that is `` not measured '', the related measurement cannot be performed until the measurement is completed and the problem analysis is completed. The process ends (step 1507).

(8) When there is no related measurement that is “not measured” in the determination in step 1507, or after the measurement adjustment process in step 1505 is completed, the recovery means implementation unit of the server monitoring manager program 105 108 is instructed to execute the recovery means for the target problem, and the measurement result column 1205 of the related problem in the problem management table 404 is set to “not measured”, and the problem handling processing here is performed. The process ends (steps 1508 and 1809).

  FIG. 16 is a flowchart for explaining the operation of the problem analysis processing in step 1504 of the flow shown in FIG. 15, and this will be described next. This problem analysis process is a process executed by the measurement result analysis unit 107 to analyze the problem range after the measurement for confirming the problem range is completed for the measurement related to the problem.

(1) When this process is started, the measurement result analysis unit 107 refers to the relation type column 1204 and the measurement result column 1205 of the problem management table 404 illustrated in FIG. It is determined whether or not there is a measurement that is “exceeding the threshold value” (step 1601).

(2) If there is a related measurement that is “exceeding the threshold value” in the measurement in which the “target application” is equal in the determination in step 1601, the measurement is performed with the related type column 1204 in the problem management table 404. With reference to the result column 1205, it is determined whether or not there is a measurement in which the “target server” is equal to “exceeding the threshold value” (step 1602).

(3) If there is a measurement in which the “target server” is equal to the “threshold exceeded” in the determination in step 1602, the related type column 1204 and the measurement result column in the problem management table 404 With reference to 1205, it is determined whether or not there is a measurement in which the “target network” is equal to “exceeding the threshold value” (step 1603).

(4) In the case where there is a measurement in which the “target network” is equal to the “threshold exceeded” in the determination in step 1603, the same server application as the problem measurement is determined in step 1601. 151, it is determined that there is a problem even with measurement from another measurement agent device 120, and it is determined at step 1602 that there is also a problem with measurement for another server application 151 operating on the same server 150 as the measurement of the problem. Since it is determined in step 1603 that there is a problem in measurement for another server application 151 operating on another server 150 in the same server network as the measurement of the problem, the scope of the problem is “same server network”. Is determined (determination 1) (step 1604).

(5) If there is no measurement in which the “target network” is equal to the “exceeding threshold” in the determination in step 1603, the same server application as the problem measurement is determined in step 1601. 151, it is determined that there is a problem even with measurement from another measurement agent device 120, and it is determined at step 1602 that there is also a problem with measurement for another server application 151 operating on the same server 150 as the measurement of the problem. Since it is determined in step 1603 that there is no problem in measurement for another server application 151 operating on another server 150 in the same server network as the measurement of the problem, the problem range is “same server 150”. It is determined that there is (determination 2) (step 1605).

(6) If there is no measurement in which the “target server” is equal to the “exceeding threshold” in the determination in step 1602, the same server application as the problem measurement is determined in step 1601. 151, it is determined that there is a problem even in measurement from another measurement agent device 120, and it is determined in step 1602 that there is no problem in measurement for another server application 151 operating on the same server 150 as the measurement of the problem. Therefore, it is determined (determination 3) that the problem range is “same server application 151” (step 1606).

(7) If there is no related measurement in which the “target application” is equal in the determination in step 1601 and there is no “exceeding threshold”, the determination in step 1601 is the same as the problem measurement. Since it is determined that there is no problem in measurement from another measurement agent device 120 for the server application 151, it is determined (determination 4) that the scope of the problem is “a combination of the server application 151 and the measurement agent device 120” (determination 4). Step 1607).

(8) After completion of the determination processing in steps 1604 to 1607, according to the determination results in steps 1604 to 1607, the measurement not included in the problem range is deleted from the problem management table 404, and the processing here ends. (Step 1608).

  FIG. 17 is a flowchart for explaining the operation of the measurement adjustment processing in step 1505 of the flow described with reference to FIG. 15. Next, this will be described. This measurement adjustment process shortens the interval of measurements with high priority in the measurement included in the problem range, so that the recovery from the problem can be detected quickly, and low priority in the measurement included in the problem range This processing is executed by the measurement result analysis unit 107 in order to reduce the load on the server application 151, the server 150, the server network, and the like included in the problem range by temporarily stopping the measurement.

  Although not described here, it may be a process of temporarily stopping the execution of all the low-priority measurements other than the high-priority measurement, and the measurement from a part or all of the measurement agent device is temporarily performed. It is also possible to stop the process.

(1) When this process is started, the measurement result analysis unit 107 sets to repeatedly execute the processes in Steps 1702 to 1705 for the measurement related to the target measurement (Step 1701, 1706).

(2) After that, the measurement setting table 402 and the problem management table 404 are referred to, and the average value of the priorities of all related measurements included in the problem range is compared with the priority of the related measurement in question. (Step 1702).

(3) Next, it is determined whether or not the priority of the related measurement as a target is higher than the average value of the priorities of all the related measurements in the comparison in Step 1702 (Step 1703).

(4) If it is determined in step 1703 that the priority of the relevant measurement as a target is higher than the average value of the priorities of all the related measurements, the relevant measurement in the measurement interval table 403 In the measurement interval column 703, a time that is 1/10 of the time defined in the measurement interval column 606 of the measurement setting table 402 is set to increase the frequency of measurement (step 1704).

(5) If it is determined in step 1703 that the priority of the relevant measurement as a target is lower than the average value of the priorities of all the related measurements, the relevant measurement in the measurement interval table 403 “Off” is set in the measurement execution on / off column 702, and the measurement is temporarily stopped (step 1705).

(6) After completing the processing of steps 1702 to 1705 repeatedly for the measurement related to the target measurement, “recovery being confirmed” is set in the status column 1202 of the problem management table 404; The internal state of the measurement result analysis unit 107 is shifted from the state of confirming the problem range to the state of confirming the recovery of the problem, and the process here is terminated (step 1707).

  FIG. 18 is a flowchart for explaining the operation of the recovery confirmation processing in step 1506 of the flow described with reference to FIG. 15. Next, this will be described. This recovery confirmation process is a process executed by the measurement result analysis unit 107 in order to confirm recovery from a problem.

(1) When this process is started, the measurement result analysis unit 107 compares the measurement result of the target measurement with the threshold value in the threshold value column 605 of the measurement setting table 402, and the comparison result is the threshold value. When the value exceeds the threshold value, “exceeding threshold” is set in the measurement result column 1205 of the problem management table 404. When the comparison result is less than the threshold value, “below threshold” is set in the measurement result column 1205. The problem management table 404 is set and updated (step 1801).

(2) Next, with reference to the measurement result column 1205 of the problem management table 404, it is determined whether or not all the recovery from the problem has been confirmed. In this determination, if there is a measurement related to “not measured” or a measurement related to “exceeding a threshold value”, the recovery confirmation process is terminated as recovery from the problem has not been confirmed ( Step 1802).

(3) If there is no measurement related to “unmeasured” and no measurement related to “exceeding threshold” in the determination in step 1802, that is, all measurements for detecting recovery from the problem. If the result does not exceed the threshold value, the recovery from the problem has been confirmed. In the measurement interval column 703, the time defined in the measurement interval column 606 of the measurement setting table 402 is set, and the measurement execution on / off column of the measurement interval table 403 for the measurement for which measurement has been temporarily stopped. In step 702, “On” is set, the measurement interval table 403 is updated, and measurement is resumed (step 1803).

(4) Thereafter, the measurement confirming the recovery from the problem and the related measurement are deleted from the problem management table 404, the problem management table 404 is updated, and the processing here is terminated (step 1804).

  FIG. 19 is a diagram showing a configuration example of the recovery means table 405 shown in FIG.

  The recovery means table 405 includes a recovery target column 1901 that defines the name or identifier of the server application 151 to be recovered from the problem, the host name or IP address of the server 150, or the network address or network name of the server network, and a problem Defines a recovery means execution destination column 1902 for defining the host name or IP address of the server resource management apparatus 140 that is a target for executing recovery means from the server, and a method for instructing the server resource management apparatus 140 to execute the recovery means And recovery means sequence 1903 to be configured.

  The recovery means registered in the recovery means column 1903 is a communication packet for instructing execution of the recovery means transmitted from the server monitoring manager program 105 to the server resource management device 140, or from the server monitoring manager program 105 by executing it. This is a command or program that can transmit a communication packet that instructs the server resource management device 140 to execute the recovery means.

  FIG. 20 is a flowchart for explaining the operation of the recovery means execution processing in the recovery means execution unit 108 of the server monitoring manager program 105 of the server monitoring apparatus 100 shown in FIG. 1, and this will be described next. This recovery means execution process is a process executed by the recovery means execution unit 108 when the measurement result analysis unit 107 is instructed to execute the recovery means including the problem determination result.

(1) When this process is started, the recovery means execution unit 108 first refers to the recovery means table 405, and the problem range notified from the measurement result analysis unit 107 is the recovery target column 1901 of the recovery means table 405. It is determined whether or not it is registered. If it is determined in this determination that the notified problem range is not registered in the recovery target column 1901 of the recovery means table 405, the recovery means execution processing here ends (step 2001).

(2) When the notified problem range is registered in the recovery target column 1901 of the recovery means table 405 in the determination of step 2001, the recovery means table 405 is referred to and registered in the recovery means execution destination column 1902. The recovery means registered in the recovery means column 1903 is executed for the server resource management apparatus 140 that is present, and the recovery means execution processing here ends (step 2002).

  Each process in the above-described embodiment of the present invention is configured by a program and can be executed by a processor (CPU) included in the present invention. The program is stored in a recording medium such as an FD, a CDROM, a DVD, or the like. It can be stored and provided, or can be provided by digital information via a network.

  FIG. 21 is a system configuration diagram for explaining the operation when the server monitoring manager program 105 of the server monitoring apparatus 100 implements recovery means on the server resource management apparatus 140. The example shown here shows the parts of the server 150 and the server resource management device 140 in the system configuration shown in FIG. 1. One server is provided as a standby server, and the server resource management device 140 is a server. This is an example in the case of using a load balancer 2120.

  The configuration example illustrated in FIG. 21 is configured by connecting a server 1 (2101), a server 2 (2102), and a server 3 (2103) to a load balancer 2120. In the server 1 (2101), AP1 (2111) and AP2 (2112) are operating as the server application 151, and in the server 2, AP3 (2113) and AP4 (2114) are operating. To do. Further, it is assumed that the server 3 (2103) is on standby as a spare.

  When the server monitoring manager program 105 detects a problem of AP1 (2111) and analyzes that the server 1 (2101) is within the problem range, the server monitoring manager program 105 sends a server to the load balancer 2120. A means for recovering 1 (2101) is executed. In the case of the described example, the server application AP1 or AP2 operating on the server 1 (2101) is moved to the waiting server 3 (2103) to operate on the problem of the server 1 (2101). It is assumed that the command to be registered is registered in the recovery means column 1903 of the recovery means table 405 shown in FIG. The server monitoring manager program 105 executes this command for the load balancer 2120, whereby the load balancer 2120 moves the AP2 of the server 1 (2101) to the server 3 (2103), and the server 3 (2103). To operate AP2. By implementing such recovery means, the load on the server 1 (2111) is reduced and the problem can be recovered.

1 is a block diagram showing the overall configuration of a server application monitoring system according to an embodiment of the present invention. It is a block diagram which shows the structural example of a server. It is a block diagram which shows the structural example of a server resource management apparatus. It is a figure which shows the table which comprises the measurement information table group stored in the memory | storage device of a server monitoring apparatus. It is a figure which shows the structural example of the measurement object table shown in FIG. It is a figure which shows the structural example of the measurement setting table shown in FIG. It is a figure which shows the structural example of the measurement space | interval table shown in FIG. It is a flowchart explaining the operation | movement of the measurement control process in the measurement control part of the server monitoring manager program of a server monitoring apparatus. It is a figure which shows the structure of the measurement result table stored in the memory | storage device of a server monitoring apparatus. It is a flowchart explaining the operation | movement of the measurement implementation process in the measurement implementation part of the server monitoring agent program in a measurement agent apparatus. It is a figure which shows the structural example of the measurement data table stored in the memory | storage device of a measurement agent apparatus. It is a figure which shows the structural example of the problem management table shown in FIG. It is a flowchart explaining operation | movement of the measurement result analysis process in the measurement result analysis part of the server monitoring manager program in a server monitoring apparatus. It is a flowchart explaining the operation | movement of the problem new registration process in step 1303 of the flow shown and demonstrated in FIG. It is a flowchart explaining the operation | movement of the problem response process in step 1304 of the flow shown and demonstrated in FIG. It is a flowchart explaining the operation | movement of the problem analysis process in step 1504 of the flow shown and demonstrated in FIG. It is a flowchart explaining the operation | movement of the measurement adjustment process in step 1505 of the flow shown and demonstrated in FIG. It is a flowchart explaining the operation | movement of the recovery confirmation process in step 1506 of the flow shown and demonstrated in FIG. It is a figure which shows the structural example of the recovery means table shown in FIG. It is a flowchart explaining the operation | movement of the recovery means implementation process in the recovery means implementation part of the server monitoring manager program of a server monitoring apparatus. It is a system block diagram explaining operation | movement when the server monitoring manager program of a server monitoring apparatus implements a recovery means to a server resource management apparatus.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 Server monitoring apparatus 101, 121 Memory 102, 124 Communication interface 103, 123 Disk interface 104, 122 Processor 105 Server monitoring manager program 106 Measurement control part 107 Measurement result analysis part 108 Recovery means implementation part 109, 127 Storage apparatus 110 Measurement information table County 111 Measurement result table 120 Measurement agent device 125 Server monitoring agent program 126 Measurement execution unit 128 Measurement data table 130 Network 140 Server resource management device 150 Monitored server 151 Server application 160 Client 161 Client application 401 Measurement target table 402 Measurement setting table 403 Measurement interval table 404 Problem management table 405 Recovery means table

Claims (5)

In a server application monitoring method in a server application monitoring system comprising a plurality of servers that execute a plurality of server applications, a plurality of measurement agent devices, and a server monitoring device that monitors the servers,
The server monitoring device
For the combination of the server application and the measurement agent device, the priority, threshold value and measurement interval are stored in advance,
According to the measurement interval, repeatedly transmits a measurement instruction of the specified server application to a plurality of the measurement agent devices,
The measurement agent device includes:
In response to a measurement instruction from the server monitoring device, measure a response time from requesting processing to the instructed server application until receiving a result of the requested processing, and measuring the measured response time, Sent to the server monitoring device,
The server monitoring device
Repeatedly receiving the response time from a plurality of the measurement agent devices,
By comparing the response time received from the measurement agent device with the threshold value, the measurement agent device that has transmitted the response time exceeding the threshold value, and the server application as a measurement target, Extract the first combination containing
Obtaining one or more second combinations related to the first combination, including the server application and the measurement agent device;
Calculating an average value of the priority of the first combination and the priority of one or more of the second combinations;
By comparing the calculated average value with the priority of the first combination and the priority of one or more of the second combinations, a third combination having a higher priority than the average value is obtained. Determining from a first combination and one or more of the second combinations;
The response time measurement instruction for the third server application included in the third combination is sent to the third combination included in the third combination at a time interval shorter than the application measurement interval corresponding to the third combination. Sent to the measurement agent device,
Determining a fourth combination having a lower priority than the average value from the first combination and the one or more second combinations;
The server characterized by temporarily stopping the transmission of the response time measurement instruction for the fourth server application included in the fourth combination to the fourth measurement agent device included in the fourth combination. Application monitoring method. The second combination including the server application and the measurement agent device related to the first combination is a combination in which a part of the first combination has an equal part. Server application monitoring method. The second combination including the server application and the measurement agent device related to the first combination is a combination having a server application, a server, or a network that the first combination has in common. The server application monitoring method according to claim 1. When the server monitoring device determines that there is a problem in the measurement in which the server application is equal in the first and second combinations, the server monitoring device determines that the problem range is a combination of the server application and the measurement agent. The server application monitoring method according to claim 2, wherein: In a server application monitoring system comprising a plurality of servers that execute a plurality of server applications, a plurality of measurement agent devices, and a server monitoring device that monitors the servers,
The server monitoring device
For the combination of the server application and the measurement agent device, the priority, threshold value and measurement interval are stored in advance,
According to the measurement interval, repeatedly transmits a measurement instruction of the specified server application to a plurality of the measurement agent devices,
The measurement agent device includes:
In response to a measurement instruction from the server monitoring device, measure a response time from requesting processing to the instructed server application until receiving a result of the requested processing, and measuring the measured response time, Sent to the server monitoring device,
The server monitoring device
Repeatedly receiving the response time from a plurality of the measurement agent devices,
By comparing the response time received from the measurement agent device with the threshold value, the measurement agent device that has transmitted the response time exceeding the threshold value, and the server application as a measurement target, Extract the first combination containing
Obtaining one or more second combinations related to the first combination, including the server application and the measurement agent device;
Calculating an average value of the priority of the first combination and the priority of one or more of the second combinations;
By comparing the calculated average value with the priority of the first combination and the priority of one or more of the second combinations, a third combination having a higher priority than the average value is obtained. Determining from a first combination and one or more of the second combinations;
The response time measurement instruction for the third server application included in the third combination is sent to the third combination included in the third combination at a time interval shorter than the application measurement interval corresponding to the third combination. Sent to the measurement agent device,
Determining a fourth combination having a lower priority than the average value from the first combination and the one or more second combinations;
The server characterized by temporarily stopping the transmission of the response time measurement instruction for the fourth server application included in the fourth combination to the fourth measurement agent device included in the fourth combination. Application monitoring system.

Images