Sliding window for alert generation

Abstract

A method for generating alerts from a periodically sampled system is disclosed. The method includes maintaining a sampling window having a predefined number of most-recently collected samples. The method further includes calculating an alert value, the alert value representing a ratio of a first value to a total number of samples in the sampling window, the first value representing a total number of samples in the sampling window having parameters exceeding a predefined sample alert threshold. The method additionally includes generating a first alert if the alert value exceeds a predefined alert threshold.

Description

BACKGROUND OF THE INVENTION

Electronic monitoring is commonly employed to monitor the state of a system. From computer system performance to home security to automobile engines, electronic monitoring is increasingly employed to generate alerts if the monitored system fails to perform satisfactorily in some manner. Monitoring can be performed on a system-wide basis or on a component of the system. In the following disclosure, system monitoring is discussed although it should be understood that such discussion applies as well to monitoring an individual component since the difference is subjective. That is, a person may consider an object a system by itself, while another may consider the same object a component in a larger system.

One of the most common ways to continuously monitor a system is by periodic sampling. In periodic sampling, the state of the monitored system is sampled on a periodic schedule. In a typical monitoring situation, some predefined alert threshold is specified. If the value of the most recently obtained sample of data exceeds the predefined alert threshold, an alert is generated. If the most recently obtained sample falls below the threshold, no alert is generated.

This method of alert generation is, however, susceptible to transient conditions. For example, if the state of the monitored system swings rapidly above and below the predefined alert threshold every sampling period or every few sampling periods, alerts are alternately sent and suppressed. If the sampling period is relatively high, e.g., many times per second or per minute, the alert signal simply jitters on and off.

SUMMARY OF INVENTION

The invention relates, in an embodiment, to a method for generating alerts from a periodically sampled system. The method includes maintaining a sampling window having a predefined number of most-recently collected samples. The method further includes calculating an alert value, the alert value representing a ratio of a first value to a total number of samples in the sampling window, the first value representing a total number of samples in the sampling window having parameters exceeding a predefined sample alert threshold. The method additionally includes generating a first alert if the alert value exceeds a predefined alert threshold.

In another embodiment, the invention relates to an article of manufacture comprising a program storage medium having computer readable code embodied therein, the computer readable code being configured to generate alerts from a periodically sampled system. There is included computer readable code for maintaining a sampling window having a predefined number of most-recently collected samples. There is further included computer readable code for calculating an alert value, the alert value representing a ratio of a first value to a total number of samples in the sampling window, the first value representing a total number of samples in the sampling window having parameters exceeding a predefined sample alert threshold. Additionally, there is included computer readable code for generating a first alert if the alert value exceeds a predefined alert threshold.

In yet another embodiment, the invention relates to a method for fine-tuning an alert signal obtained from a periodically sampled system. There is included providing an alert signal generator. The alert signal generator includes first logic for maintaining a sampling window at a constant window size, the sampling window containing most-recently collected samples. The alert signal generator includes second logic for calculating an alert value, the alert value representing a ratio of a first value to a total number of samples in the sampling window, the first value representing a total number of samples in the sampling window having parameters exceeding a sample alert threshold. The alert generator includes third logic for generating a first alert if the alert value exceeds an alert threshold. There is also included providing a control panel having a plurality of user-manipulable controls. The plurality of user-manipulable controls include at least one of a first control for varying the window size, a second control for varying the sample alert threshold, and a third control for varying the alert threshold.

These and other features of the present invention will be described in more detail below in the detailed description of various embodiments the invention and in conjunction with the following figures.

BRIEF DESCRIPTION OF THE DRAWINGS

The present invention is illustrated by way of example, and not by way of limitation, in the figures of the accompanying drawings and in which like reference numerals refer to similar elements and in which:

FIG. 1 illustrates, in accordance with an embodiment of the invention, a sliding window having a plurality of samples.

FIG. 2 illustrates, in accordance with an embodiment of the present invention, the window of FIG. 1 after another sample is obtained.

FIG. 3 illustrates, in accordance with an embodiment of the present invention, the window of FIG. 2 after another sample is obtained.

FIG. 4 illustrates, in accordance with an embodiment of the present invention, the window of FIG. 3 after another sample is obtained.

FIG. 5 illustrates, in accordance with an embodiment of the present invention, the window of FIG. 4 after another sample is obtained.

FIG. 6 illustrates, in accordance with an embodiment of the present invention, the window of FIG. 5 after another sample is obtained.

FIG. 7 illustrates, in accordance with an embodiment of the present invention, the window of FIG. 6 after another sample is obtained.

FIG. 8 illustrates, in accordance with an embodiment of the present invention, the window of FIG. 7 after another sample is obtained.

FIG. 9 shows, in accordance with an embodiment of the invention, the sliding window technique for generating alerts.

DETAILED DESCRIPTION OF VARIOUS EMBODIMENTS

The present invention will now be described in detail with reference to various embodiments thereof as illustrated in the accompanying drawings. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention. It will be apparent, however, to one skilled in the art, that the present invention may be practiced without some or all of these specific details. In other instances, well known process steps and/or structures have not been described in detail in order to not unnecessarily obscure the present invention.

Various embodiments are described hereinbelow, including methods and techniques. It should be kept in mind that the invention may also cover articles of manufacture that includes a computer readable medium on which computer-readable instructions for carrying out embodiments of the inventive technique are stored. The computer readable medium may include, for example, semiconductor, magnetic, opto-magnetic, optical, or other forms of computer readable medium for storing computer readable code. Further, the invention may also cover apparatuses for practicing embodiments of the invention. Such apparatus may include circuits, dedicated and/or programmable, to carry out tasks pertaining to embodiments of the invention. Examples of such apparatus include a general purpose computer and/or a dedicated computing device when appropriately programmed and may include a combination of a computer/computing device and dedicated/programmable circuits adapted for the various tasks pertaining to embodiments of the invention.

In accordance with embodiments of the present invention, there is provided a technique for substantially eliminating alert jitters due to transient conditions in the periodically sampled monitored system. Embodiments of the invention provide a persistent sliding window having a predefined number of most-recently collected samples. The predefined number of most-recently collected samples may be predefined by the system or may be definable by the user.

As a sample is collected, this most-recently collected sample is added to the sliding window, and the oldest sample in the sliding window is discarded. Thus, the size of the sliding window remains constant despite the fact that its membership changes as the samples are collected.

Periodically or upon the collection of a new sample, an alert value is calculated. This alert value calculation may be explained using an example. Suppose there are ten samples in the sliding window. As new samples are collected, they are added to the sliding window and the oldest samples are discarded, thereby keeping the sliding window membership at 10.

Suppose each sample contains a parameter having a value between 0 and 1, with 0.5 representing the predefined sample alert threshold (SAT). The predefined sample alert threshold (SAT) represents a system-defined or user-definable value that is employed to ascertain whether a sample collected reflects an event that potentially raises an alert or is simply a normal state of operation. In the case where the predefined SAT is 0.5, if a sample parameter is greater than 0.5, that sample is deemed a potential alert-causing sample (PACS). The borderline case wherein the sampled parameter is equal to the SAT (e.g., 0.5 in the current example), the sample may be deemed a PACS or a non-PACS, depending on how the alert generating algorithm is configured.

The calculated alert value (CAV) represents the ratio of PACS to the total number of samples in the sliding window. If there are four samples whose parameters exceed the predefined SAT value of 0.5, the number of PACS will be 4. In this case, the calculated alert value will be 4 divided by 10 or 40%.

Furthermore, there is provided a predefined alert threshold (AT), which may be a system-defined or a user-definable value. If the calculated alert value (CAV) exceeds the predefined alert threshold (AT), an alert is generated. On the other hand, if the CAV is below the predefined alert threshold (AT), the alert is not generated. In the borderline case wherein the CAV equals the predefined alert threshold (AT), an alert may be generated or may be suppressed depending on how the alert generating algorithm is configured.

Since the alert is generated based on the calculated alert value associated with the sliding window instead of being based solely on the value of the most recently collected sample, jitters in the alert signal are substantially eliminated. Further, the user may be furnished with controls in some embodiments to tune the alerting system to generate alerts based on user-specified thresholds. These controls may be furnished on a computer screen as part of a control panel to allow the user to fine-tune the alert signal, for example. As an example, if jitter is deemed excessive in the alert signal received, the user may increase the number of samples in the sliding window, thereby lessening the impact of transients. As another example, if additional sensitivity is desired, the user may decrease the number of samples in the sliding window and/or lowering the sample alert threshold (SAT) and/or lowering the predefined alert threshold (AT). One skilled in the art would readily appreciate that providing these user controls to allow the user to fine-tune the sliding window-based alert signal represents a level of refinement not available in the prior art.

The features and advantages of the present invention may be better understood with reference to the figures and drawings below. Suppose an internet merchant wishes to be alerted if the time it takes to serve up a webpage from his internet merchandising server is greater than 10 seconds. In FIG. 1, each vertical line represents a sample of data collected every 5 minutes from a data probe that simulates the loading of the webpage of interest. Five initial samples (S1, S2, S3, S4, and S5) are shown. The sliding window is shown to be 20 minutes, spanning five samples S1–S5.

In FIG. 1, the samples having parameters exceeding the predefined sample alert threshold (SAT) are shown with large arrows. In this example, for samples S2 and S3, the time taken to serve up the webpage of interest is greater than the sample alert threshold value of 10 seconds. Accordingly, arrows A2 and A3 are shown above samples S2 and S3.

In this case, the calculated alert value (CAV) is ⅖ or 40% since two out of the five samples in the sliding window exceed the predefined sample alert threshold (SAT). If the predefined alert threshold is 50%, i.e., the CAV needs to be 50% or higher to generate an alert, no alert would be generated in the case of FIG. 1.

In FIG. 2, another sample S6 is collected. Sample S1 is dropped from the sliding window to keep the size of the sliding window constant. In this case, sample S6 happens to exceed the predefined sample alert threshold (SAT) and is thus shown with a large white arrow A6 in FIG. 2. The calculated alert value (CAV) is 60% since 3 out of the 5 samples in the sliding window exceed the predefined sample alert threshold (SAT). Furthermore, since the CAV exceeds the predefined alert threshold (AT) of 50%, an alert is generated. The alert is furnished to the user to indicate a change from the non-alert state of FIG. 1 to the alert state of FIG. 2.

In FIG. 3, another sample S7 is collected. Sample S2 is dropped from the sliding window to keep the size of the sliding window constant. In this case, sample S7 happens to be below the predefined sample alert threshold (SAT). The calculated alert value (CAV) is 40% since only 2 out of the 5 samples in the sliding window exceed the predefined sample alert threshold (SAT). Furthermore, since the CAV does not exceed the predefined alert threshold (AT) of 50%, no alert is generated.

In an embodiment, a notification may be generated if the monitored system moves from an alert state (such as in FIG. 2) to a non-alert state (such as in FIG. 3). The notification may be generated in this case to let the user know that the alert state no longer exists, for example. In some systems, the user may be given a choice to set a preference setting indicating whether such a notification is desired when the system moves from an alert state in one sampling period to a non-alert state in the immediately succeeding sampling period.

In FIG. 4, another sample S8 is collected. Sample S3 is dropped from the sliding window to keep the size of the sliding window constant. In this case, sample S8 happens to exceed the predefined sample alert threshold (SAT) and is thus shown with a large white arrow A8 in FIG. 4. The calculated alert value (CAV) is 40% since only 2 out of the 5 samples in the sliding window exceed the predefined sample alert threshold (SAT). Furthermore, since the CAV does not exceed the predefined alert threshold (AT) of 50%, no alert is generated.

In FIG. 5, another sample S9 is collected. Sample S4 is dropped from the sliding window to keep the size of the sliding window constant. In this case, sample S9 happens to exceed the predefined sample alert threshold (SAT) and is thus shown with a large white arrow A9 in FIG. 5. The calculated alert value (CAV) is 60% since 3 out of the 5 samples in the sliding window exceed the predefined sample alert threshold (SAT). Furthermore, since the CAV exceeds the predefined alert threshold (AT) of 50%, an alert is generated. The alert is furnished to the user to indicate a change from the non-alert state of FIG. 4 to the alert state of FIG. 5.

In FIG. 6, another sample S10 is collected. Sample S5 is dropped from the sliding window to keep the size of the sliding window constant. In this case, sample S10 happens to exceed the predefined sample alert threshold (SAT) and is thus shown with a large white arrow A10 in FIG. 6. The calculated alert value (CAV) is 80% since 4 out of the 5 samples in the sliding window exceed the predefined sample alert threshold (SAT). Furthermore, since the CAV exceeds the predefined alert threshold (AT) of 50%, an alert is generated.

In an embodiment, a new alert is furnished to the user when moving from the alert state of FIG. 5 to the alert state of FIG. 6. In another embodiment, the new alert is generated but not furnished to the user if the user chooses not to be bothered with a new alert if the state did not change (i.e., simply moving from one alert state to another alert state). In an embodiment, however, the user may be alerted if a different alert message is generated. For example, a sample may exceed the predefined sample alert threshold (SAT) and may be classified as one of several categories of severity (e.g., warning, minor, major, critical). The user may elect, via a preference setting, to be notified not only with an alert but also with the highest severity level experienced by any of the samples in the sliding window. In this case, if the highest severity level changes when moving from one sampling period to the next, the user may elect to be notified of such change, and the alert system would appropriately notify the user when such change occurs.

In FIG. 7, another sample S11 is collected. Sample S6 is dropped from the sliding window to keep the size of the sliding window constant. In this case, sample S11 happens to be below the predefined sample alert threshold (SAT) and is thus not shown with a large white arrow in FIG. 7. The calculated alert value (CAV) is 60% since 3 out of the 5 samples in the sliding window exceed the predefined sample alert threshold (SAT). Furthermore, since the CAV exceeds the predefined alert threshold (AT) of 50%, an alert is generated.

In FIG. 8, another sample S12 is collected. Sample S7 is dropped from the sliding window to keep the size of the sliding window constant. In this case, sample S12 happens to be below the predefined sample alert threshold (SAT) and is thus not shown with a large white arrow in FIG. 8. The calculated alert value (CAV) is 60% since 3 out of the 5 samples in the sliding window exceed the predefined sample alert threshold (SAT). Furthermore, since the CAV exceeds the predefined alert threshold (AT) of 50%, an alert is generated.

In an embodiment, the calculated alert value is not computed until the sliding window is filled up. In the context of the current example, no calculation would be done until at least five samples are collected. In another embodiment, calculation is performed with the uncollected samples counted as “non-alert” samples. Thus, if only three samples are collected and all exceed the predefined sample alert threshold (SAT), an alert would be generated since the predefined alert threshold (AT) of 50% would have been exceeded irrespective whether the last two samples of the sliding window, which are yet to be collected at that point, turn out to be “non-alert” samples or “alert” samples.

FIG. 9 shows, in accordance with an embodiment of the invention, the sliding window technique for generating alerts. In step 902, the size of the sliding window is maintained as new samples are added. As mentioned, the size of the sliding window may be maintained in an embodiment by adding the most recently collected sample to the sliding window and by discarding the oldest sample in the sliding window. In step 904, the alert value is calculated by, in an embodiment, dividing the number of samples having parameters exceeding the predefined sample alert threshold (SAT) by the number of samples in the sliding window.

In step 906, the calculated alert value (CAV) is compared against a predefined alert threshold (AT). If the calculated alert value (CAV) exceeds the predefined alert threshold (AT), an alert is generated (step 908). The method returns from step 908 to step 902 wherein the sliding window membership is changed when a new sample is collected. On the other hand, if the calculated alert value (CAV) is below the predefined alert threshold (AT), the method returns from step 906 to step 902 wherein the sliding window membership is changed when a new sample is collected.

In an embodiment, the sliding window alarming technique is employed as a part of an Internet Service monitoring system, such as the OpenView Internet Services™ or OVIS™ product from the Hewlett-Packard Company of Palo Alto, Calif. The sample may be obtained by periodically sending a synthetic transaction from probes to a target. The probe may create a HTTP request, which is then sent to the HTTP server. The manner in which the server responds to the probe's request may be used to calculate a variety of metrics, such as server availability and server response time.

Table 1 shows, in an embodiment of the present invention, the various probes that may be employed in such an Internet service monitoring system.

TABLE 1
DHCP (Dynamic Host Configuration Protocol)
DIAL (Dial-Up Networking Service)
DNS (Domain Name System)
Exchange (MAPI)
FTP (File Transfer Protocol)
HTTP (Hypertext Transfer Protocol
HTTPS (Hypertext Transfer Protocol Secure)
HTTP_TRANS (Web Transaction Recorder)
ICMP (Internet Control Message Protocol-Ping)
IMAP4 (Internet Message Access Protocol)
LDAP (Lightweight Directory Access Protocol)
MAILROUNDTRIP (Mail Round Trip)
NNTP (Network News Transfer Protocol)
NTP (Network Time Protocol)
ODBC (Open Database Connectivity)
POP3 (Post Office Protocol 3)
RADIUS (Remote Authentication Dial In User Service)
SAP Basis
Script (generic script)
SMS (Short Message Service)
SMTP (Simple Mail Transfer Protocol)
SOAP (Simple Object Access Protocol)
STREAM_MEDIA (Streaming Media)
TCP (Transmission Control Protocol)
WAP (Wireless Application Protocol)

Table 2 shows, in an embodiment of the present invention, the various metrics that may be measured from the server's response to the request sent via the probe.

TABLE 2
ANSWER_DNS
AUTH_TIME
AVAILABILITY
BAUD_RATE
BROKEN_LINKS
BWDLOSS
CONGESTION
CONNECT_TIME
CUSTOMER_NAME
DATA_TRANS_BYTES
DNS_SETUP_TIME
FAILED_TRANSACTION_RATE
FWDLOSS
GROUP_TIME
HTTP_STATUS
HTTP_TIME
LATENCY
LEASE_TIME
MAX_RESPONSE
MIN_RESPONSE
NUM_ENTRIES
OFFER_TIME
PACKETS_RECEIVED
PACKET_LOSS
PORT_TIME
RAS_CONNECT_STATUS
READ_TIME
REQUESTS
RESPONSE_TIME
RESPONSE_TIME_VIOLATION_COUNT
RESPONSE_TIME_VIOLATION_PERCENTAGE
SEND_TIME
SERVER_RESPONSE_TIME
SERVER_RESP_TIME
SERVICE_NAME
SETUP_TIME
STREAM_SETUP_TIME
TEAR_DOWN_TIME
TERMINATION_STATUS
TOTAL_CONNECTION_TIME
TRANSACTION_RATE
TRANSACTION_SIZE
TRANSACT_AVG
TRANSFER_BYTES
TRANSFER_TIME
TRANSFER_TPUT
TRANS_BYTES

Table 3 shows, in an implementation of the present invention, various probe types and associated metrics that may be collected along with their descriptions.

TABLE 3
                 Metrics (in italic; Metric 1–8 are labels)
Probe type       and Description
DHCP             AVAILABILITY - If a measurement could not be
                 retrieved a 0 is set
                 otherwise availability is set to 1
                 RESPONSE_TIME - Total response time for the
                 DHCP service (Setup
                 Time + Transfer Time)
                 SETUP_TIME - Time to resolve address and
                 establish the connection if
                 host is specified
                 TRANSFER_TPUT - Transfer bytes/Transfer
                 Time in kbytes/sec
                 OFFER_TIME - Metric 1 - Time to first offer
                 from server
                 LEASE_TIME - Metric 2 - Time to lease offered
                 IP address
                 TRANSFER_TIME - Metric 5 - Time to complete
                 entire transaction
                 (discover, offer, request, acknowledge and
                 release)
                 TRANS_BYTES - Metric 6 - The number of bytes
                 transferred
DIAL             AVAILABILITY - If a measurement could not be
                 retrieved a 0 is set
                 otherwise availability is set to 1
                 RESPONSE_TIME - Time taken to establish PPP
                 connection
                 RAS_CONNECT_STATUS - Metric 1 - Error
                 returned by RAS Dial Will
                 be 0 for successful connection
                 BAUD_RATE - Metric 2 - Baud Rate - Transfer
                 rate as reported by the
                 modem
                 TOTAL_CONNECTION_TIME - Metric 3 - Total
                 time connected
                 TERMINATION_STATUS - Metric 4 - True (1)
                 for abnormal termination
                 of connection, otherwise false (0)
DNS              AVAILABILITY - If a measurement could not be
                 retrieved a 0 is set
                 otherwise availability is set to 1
                 RESPONSE_TIME - Execution time of the query
                 to a hostname/IP
                 address
                 ANSWER_DNS - Metric 1 - Answer DNS is set to
                 0 if the hostname cannot
                 be resolved, and 1 if it can In either case
                 Availability will be 1 (or true)
                 because the server is doing
Exchange         AVAILABILITY - If a measurement could not be
                 retrieved a 0 is set,
                 otherwise availability is set to 1
                 SETUP_TIME - Time to log in to the Exchange
                 Server and resolve the
                 name
                 RESPONSE_TIME - Total response time of the
                 Exchange service Setup
                 Time + time to read all messages and mark the
                 OVIS ones for delete
FTP              AVAILABILITY - If a measurement could not be
                 retrieved a 0 is set,
                 otherwise availability is set to 1
                 RESPONSE_TIME - Total response time of the
                 FTP request (DNS Setup
                 Time + Connect Time + Server Response Time +
                 Authentication Time +
                 Port Time + Transfer Time)
                 SETUP_TIME - Time to resolve address and
                 establish the connection
                 TRANSFER_TPUT - Transfer bytes/Transfer
                 Time in kbytes/sec
                 DNS_SETUP_TIME - Metric 1 - Time to resolve
                 hostname through DNS
                 CONNECT_TIME - Metric 2 - Time to perform
                 connect to FTP server
                 SERVER_RESP_TIME - Metric 3 - Time it takes
                 to receive the FTP start
                 header (220)
                 AUTH_TIME - Metric 4 - Time to authenticate
                 user (time to send
                 username/password and receive response)
                 PORT_TIME - Metric 5 - Time to send the client
                 connection ports to the
                 FTP server
                 TRANSFER_TIME - Metric 6 - Overall time to
                 receive data on the data
                 connection
                 DATA_TRANS_BYTES - Metric 7 - The number
                 of bytes transferred
HTTP/HTTPS       AVAILABILITY - If a measurement could not be
                 retrieved a 0 is set,
                 otherwise availability is set to 1
                 RESPONSE_TIME - Total response time for the
                 web page (or secure web
                 page) access (DNS Setup Time + Connect
                 Time + Server Response Time +
                 Transfer Time)
                 SETUP_TIME - Time to resolve address and
                 establish the connection
                 TRANSFER_TPUT - Transfer bytes/Transfer
                 Time in kbytes/sec
                 DNS_SETUP_TIME - Metric 1 - Time to resolve
                 hostname through DNS
                 CONNECT_TIME - Metric 2 - Time to connect to
                 HTTP/S server or proxy
                 SERVER_RESP_TIME - Metric 3 - Time it takes
                 to send HTTP/S Get
                 request and receive first response packet
                 TRANSFER_TIME - Metric 4 - Time it took to
                 send request and receive all
                 reply packets
                 TRANS_BYTES - Metric 5 - The number of bytes
                 transferred
                 HTTP_STATUS - Metric 6 - HTTP/S status code
                 REQUESTS - Metric 7 - Number of HTTP/S
                 requests For example, if the
                 page was redirected or embedded objects are
                 downloaded
                 BROKEN_LINKS - Metric 8 - Number of
                 embedded objects that couldn't
                 be downloaded (eg, URL not found)
HTTP_TRANS       URL/Navigation Point Mode
                 AVAILABILITY - If a measurement could not be
                 retrieved a 0 is set,
                 otherwise availability is set to 1
                 RESPONSE_TIME -
                 Step: Total response time for the web page access
                 (DNS Setup Time +
                 Connect Time + Server Response Time +
                 Transfer Time)
                 Transaction: Total response time for all steps
                 SETUP_TIME -
                 Step: Time to resolve address and establish the
                 connection
                 Transaction: Total setup time for all steps
                 TRANSFER_TPUT -
                 Step: Transfer bytes/Transfer Time in kbytes/sec
                 Transaction: Total transfer throughput for the
                 transaction
                 DNS_SETUP_TIME - Metric 1 -
                 Step: Time to resolve hostname through DNS
                 Transaction: Total DNS setup time for all steps
                 CONNECT_TIME - Metric 2 -
                 Step: Time to connect to HTTP/S server or proxy
                 Transaction: Total connect time for all steps
                 SERVER_RESP_TIME - Metric 3 -
                 Step: Time it takes to send HTTP Get request and
                 receive first response
                 packet
                 Transaction: Total server response time for all
                 steps
                 TRANSFER_TIME - Metric 4 -
                 Step: Time it took to send request and receive all
                 reply packets
                 Transaction: Total transfer time for all steps
                 TRANSFER_BYTES - Metric 5 -
                 Step: The number of bytes transferred
                 Transaction: Total transfer bytes for all steps
                 HTTP_STATUS - Metric 6 -
                 Step: HTTP status code
                 Transaction: HTTP status code of the last step
                 REQUESTS - Metric 7 -
                 Step: Number of HTTP requests For example, if
                 the page was redirected
                 or embedded objects are downloaded
                 Transaction: Total requests for all steps
                 BROKEN_LINKS - Metric 8 -
                 Step: Number of embedded objects that couldn't
                 be downloaded (eg, URL
                 not found)
                 Transaction: Total broken links for all steps
ICMP             AVAILABILITY - If a measurement could not be
                 retrieved a 0 is set,
                 otherwise availability is set to 1
                 RESPONSE_TIME - Response time is the
                 average roundtrip time for all
                 ICMP packets
                 TRANSFER_TPUT - Transfer bytes/Transfer
                 Time in kbytes/sec
                 MIN_RESPONSE - Metric 1 - Minimum
                 roundtrip time of all ICMP
                 packets
                 MAX_RESPONSE - Metric 2 - Maximum
                 roundtrip time of all ICMP
                 packets
                 PACKET_LOSS - Metric 3 - Number of packets
                 lost
IMAP4            AVAILABILITY - If a measurement could not be
                 retrieved a 0 is set,
                 otherwise availability is set to 1
                 RESPONSE_TIME - Total response time for the
                 IMAP4 service (Setup
                 Time + Connection Time + Server Response
                 Time + Authentication Time +
                 Transfer Time)
                 SETUP              —              TIME - Time to resolve address and
                 establish the connection
                 TRANSFER_TPUT - Transfer bytes/Transfer
                 Time in kbytes/sec
                 DNS_SETUP_TIME - Metric 1 - Time to resolve
                 hostname through DNS
                 CONNECT_TIME - Metric 2 - Time to connect to
                 IMAP server
                 SERVER_RESP_TIME - Metric 3 - Time for
                 IMAP server to respond
                 AUTH_TIME - Metric 4 - Time to authenticate
                 user (time to send
                 username/password and receive response)
                 TRANSFER_TIME - Metric 5 - Overall time it
                 took for the data transfer
                 only
                 DATA_TRANS_BYTES - Metric 6 - The number
                 of bytes transferred
LDAP             AVAILABILITY - If a measurement could not be
                 retrieved a 0 is set,
                 otherwise availability is set to 1
                 RESPONSE_TIME - Total response time for the
                 LDAP service (Setup
                 Time + Data Transfer Time)
                 TRANSFER_TPUT - Transfer bytes/Transfer
                 Time in kbytes/sec
                 DNS_SETUP_TIME - Metric 1 - Time to resolve
                 hostname through DNS
                 NUM_ENTRIES - Metric 2 - Number of returned
                 entries
                 CONNECT_TIME - Metric 3 - Time to connect to
                 LDAP server
                 TRANSFER_TIME - Metric 4 - Overall time it
                 took for the data transfer
                 only
                 TRANS_BYTES - Metric 5 - The number of bytes
                 transferred
MAILROUNDTRIP    AVAILABILITY - If a measurement could not be
                 retrieved a 0 is logged,
                 otherwise availability is set to 1
                 SETUP_TIME - Time to resolve address and
                 establish the connection
                 RESPONSE_TIME - Total response time for the
                 SMTP mail send + the
                 POP/IMAP receive
                 TRANSFER_TPUT - Transfer bytes/Transfer
                 Time in kbytes/sec
NNTP             AVAILABILITY - If a measurement could not be
                 retrieved a 0 is set,
                 otherwise availability is set to 1
                 RESPONSE_TIME - Total response time for
                 NNTP (DNS Setup Time +
                 Connect Time + Server Response Time +
                 Authentication Time + Group
                 Time + Read Time + Tear Down Time)
                 SETUP_TIME - Time to resolve address and
                 establish the connection
                 TRANSFER_TPUT - Transfer bytes/Transfer
                 Time in kbytes/sec
                 DNS_SETUP_TIME - Metric 1 - Time to resolve
                 hostname through DNS
                 CONNECT_TIME - Metric 2 - Time to connect to
                 NNTP server
                 SERVER_RESP_TIME - Metric 3 - Overall time
                 to read the file (receive
                 data on the data connection)
                 AUTH_TIME - Metric 4 - Time to authenticate
                 user (time to send
                 username/password and receive response)
                 GROUP_TIME - Metric 5 - Time to select
                 newsgroup and get request
                 overview of last 100 articles
                 READ_TIME - Metric 6 - Time to read articles
                 with the overall size of
                 10000 bytes
                 TEAR_DOWN_TIME - Metric 7 - Overall time to
                 send the QUIT request
                 and receive the response
                 DATA_TRANS_TIME - Metric 8 - The number of
                 bytes transferred
NTP              AVAILABILITY - If a measurement could not be
                 retrieved a 0 is set,
                 otherwise availability is set to 1
                 RESPONSE_TIME - Total response time for the
                 NTP service (Setup Time +
                 Transfer Time)
                 SETUP_TIME - Time to resolve address and
                 establish the connection
                 TRANSFER_TPUT - Transfer bytes/Transfer
                 Time in kbytes/sec
                 DATA_TRANS_BYTES - Metric 5 - The number
                 of bytes transferred
                 TRANSFER_TIME - Metric 6 - Overall time it
                 took for the data transfer
                 only
ODBC             AVAILABILITY - If a measurement could not be
                 retrieved a 0 is set,
                 otherwise availability is set to 1
                 RESPONSE_TIME - Total response time for the
                 ODBC service
                 SETUP_TIME - Time to setup database
                 connection handles
                 TRANSFER_TPUT - Transfer bytes/Transfer
                 Time in kbytes/sec
                 CONNECT_TIME - Metric 1 - Time to connect to
                 database
                 SERVER_RESP_TIME - Metric 2 - Time to
                 respond to the SQL statement
                 TRANSFER_TIME - Metric 3 - Overall time it
                 took for the data transfer
                 DATA_TRANS_BYTES - Metric 4 - The number
                 of bytes transferred
OVTA Application AVAILABILITY - WEBAPP service type only The
(COMAPP,         ratio of availability
JMSAPP, RMIAPP   probe requests that failed, to the total attempts
SOAPAPP,         during the last interval
WEBAPP)          RESPONSE_TIME - The average response time
                 of the successfully
                 completed transactions during the interval
                 TRANSACTION_RATE - Metric 1 - Total number
                 of completed
                 transactions per second over the last interval
                 RESPONSE_TIME_VIOLATION_COUNT -
                 Metric 2 - Number of
                 successfully completed transactions in the last
                 interval whose measured
                 response time exceeded the response time
                 threshold configured in OVTA
                 RESPONSE_TIME_VIOLATION_PERCENTAGE
                 - Metric 3 - Percent of
                 successfully completed transactions in the last
                 interval whose measured
                 response time exceeded the response time
                 threshold configured in OVTA
                 TRANSACTION_SIZE - Metric 4 - WEBAPP
                 service type only The
                 average size of the successfully completed
                 transactions The size varies
                 depending on the type of application and the type
                 of transaction
                 For transactions measured in the browser using
                 an OVTA Browser Client
                 Monitor, the size of these transactions is the size
                 of the downloaded page
                 plus all embedded content and images For
                 transactions measured at the
                 Web or Application server using an OVTA Web
                 Server Monitor, the size of
                 these transactions is the size of the downloaded
                 page as reported in the
                 Content-Length HTTP header This is the size of
                 the page itself and does
                 not include embedded images Moreover, some
                 web-based applications will
                 not set the Content-Length field; therefore, this
                 metric will be 0 for
                 transactions in these types of applications
                 FAILED_TRANSACTION_RATE - Metric 5 -
                 Total number of failed
                 transactions per second over the last interval
POP3             AVAILABILITY - If a measurement could not be
                 retrieved a 0 is set,
                 otherwise availability is set to 1
                 RESPONSE_TIME - Total response time for the
                 POP3 Mail delivery
                 (DNS Setup Time + Connect Time + Server
                 Response Time +
                 Authentication Time + Data Transfer Time)
                 SETUP_TIME - Time to resolve address and
                 establish the connection
                 TRANSFER_TPUT - Transfer bytes/Transfer
                 Time in kbytes/sec
                 DNS_SETUP_TIME - Metric 1 - Time to resolve
                 hostname through DNS
                 CONNECT TIME - Metric 2 - Time to connect to
                 POP3 server
                 SERVER_RESP_TIME - Metric 3 - Time it takes
                 to receive the POP3 start
                 header (+OK)
                 AUTH_TIME - Metric 4 - Time to authenticate
                 user (time to send
                 username/password and receive response)
                 TRANSFER_TIME - Metric 5 - Overall time to
                 read all messages in the
                 mailbox and delete the IOPS test messages
                 DATA_TRANS_BYTES - Metric 6 - The number
                 of bytes transferred
RADIUS           AVAILABILITY - If a measurement could not be
                 retrieved a 0 is set,
                 otherwise availability is set to 1 If the server is
                 successfully contacted
                 but returns an Access-Reject packet (because of a
                 bad password, secret,
                 etc) the Availability will be 0
                 RESPONSE_TIME - Total response time for the
                 RADIUS service (DNS
                 Setup Time + Data Transfer Time)
                 SETUP_TIME - Time to resolve address and
                 make connection
                 TRANSFER_TPUT - Transfer bytes/Transfer
                 Time in kbytes/sec
                 TRANSFER_TIME - Metric 4 - Overall time it
                 took for the data transfer
                 only
                 DATA_TRANS_BYTES - Metric 5 - The number
                 of bytes transferred
SAP              AVAILABILITY - If a measurement could not be
                 retrieved a 0 is set,
                 otherwise availability is set to 1 Availability
                 requires both a successful
                 connection and a successful RFC call
                 RESPONSE_TIME - Total response time for the
                 SAP service Setup Time +
                 Completion which is the time to complete a
                 successful RFC call
                 (including logon check and logout)
                 SETUP_TIME - Time to get a successful
                 connection with the RFC server
Script           AVAILABILITY - If a measurement could not be
                 retrieved a 0 is set,
                 otherwise availability is set to 1
                 RESPONSE_TIME - Total time running the script
                 Or the total response
                 time imported from the Result File script
SMS              AVAILABILITY - If a measurement could not be
                 retrieved a 0 is set,
                 otherwise availability is set to 1
                 RESPONSE_TIME - Total response time for the
                 SMS service
SMTP             AVAILABILITY - If a measurement could not be
                 retrieved a 0 is set,
                 otherwise availability is set to 1
                 RESPONSE_TIME - Total response time for the
                 SMTP mail request (DNS
                 Setup Time + Connect Time + Server Response
                 Time + Transfer Time +
                 Tear Down Time)
                 SETUP_TIME - Time to resolve address and
                 establish the connection
                 TRANSFER_TPUT - Transfer bytes/Transfer
                 Time in kbytes/sec
                 DNS_SETUP_TIME - Metric 1 - Time to resolve
                 hostname through DNS
                 CONNECT_TIME - Metric 2 - Time to connect to
                 SMTP server
                 SERVER_RESP_TIME - Metric 3 - Time it takes
                 to receive the SMTP
                 start header (220)
                 TRANSFER_TIME - Metric 4 - Overall time to
                 transfer the mail request
                 (including SMTP responses to the requests such
                 as MAIL FROM:, RCPT
                 TO: DATA, QUIT
                 TRANS_BYTES - Metric 5 - The number of bytes
                 transferred
                 TEAR_DOWN_TIME - Metric 6 - Overall time to
                 send the QUIT request
                 and receive the response
SOAP             AVAILABILITY - If a measurement could not be
                 retrieved a 0 is set,
                 otherwise availability is set to 1
                 RESPONSE_TIME - Total response time for the
                 web page access (DNS
                 Setup Time + Connect Time + Server Response
                 Time + Transfer Time)
                 SETUP              —              TIME - Time to resolve address and
                 establish the connection
                 TRANSFER_TPUT - Transfer bytes/Transfer
                 Time in kbytes/sec
                 DNS_SETUP_TIME - Metric 1 - Time to resolve
                 hostname through DNS
                 CONNECT_TIME - Metric 2 - Time to connect to
                 SOAP server or proxy
                 SERVER_RESP_TIME - Metric 3 - Time it takes
                 to send HTTP Get
                 request and receive first response packet
                 TRANSFER_TIME - Metric 4 - Time it took to
                 send request and receive all
                 reply packets
                 TRANS_BYTES - Metric 5 - The number of bytes
                 transferred
                 HTTP_STATUS - Metric 6 - HTTP/S status code
                 REQUESTS - Metric 7 - Number of HTTP
                 requests For example, if the
                 page was redirected or embedded objects are
                 downloaded
                 BROKEN_LINKS - Metric 8 - Number of
                 embedded objects that couldn't
                 be downloaded (eg, URL not found)
STREAM_MEDIA     AVAILABILITY - If a measurement could not be
                 retrieved a 0 is set,
                 otherwise availability is set to 1
                 RESPONSE_TIME - Total response time for the
                 Streaming Media service
                 (Setup Time + Connect Time + Server Response
                 Time + Transfer Time)
                 SETUP_TIME - Time to resolve address
                 TRANSFER_TPUT -The average bandwidth used
                 in data transfer in
                 Kbytes/sec
                 CONNECT_TIME - Metric 1 - The time to
                 connect to the server If a proxy
                 is used then this is the time it takes to connect to
                 the proxy
                 SERVER_RESP_TIME - Metric 2 - The time it
                 takes for the server to
                 start sending packets This includes the set up time
                 for the various
                 protocols
                 TRANSFER_TIME - Metric 3 - The time it takes
                 to transfer the data
                 PACKETS_RECEIVED - Metric 4 - Total number
                 of packets received
                 PACKET_LOSS - Metric 5 - The percentage of
                 packets lost
                 LATENCY - Metric 6 - The latency in data
                 transfer in seconds The server
                 responds at set intervals so after a request is sent
                 there may be some wait
                 time before the next interval
                 CONGESTION - Metric 7 - The percentage of
                 time spent in buffering data
                 vs the total time for playing the streams This
                 excludes the initial
                 buffering time
                 STREAM_SETUP_TIME - Metric 8 - The initial
                 buffering time, before the
                 stream actually starts playing on the client
TCP              AVAILABILITY - If a measurement could not be
                 retrieved a 0 is set,
                 otherwise availability is set to 1
                 RESPONSE_TIME - Total response time for the
                 TCP service (DNS Setup
                 Time + Connect Time)
                 SETUP              —              TIME - Time to resolve address and
                 establish the connection
                 TRANSFER_TPUT - Transfer bytes/Transfer
                 Time in kbytes/sec
                 DNS_SETUP_TIME - Metric 1 - Time to resolve
                 hostname through DNS
                 CONNECT              —              TIME - Metric 2 - Time to perform
                 connect to resolved IP
                 address
WAP              AVAILABILITY - If a measurement could not be
                 retrieved a 0 is set,
                 otherwise availability is set to 1
                 RESPONSE_TIME - Total response time for the
                 WAP service (DNS Setup
                 Time + Transfer Time)
                 TRANSFER_TPUT - Transfer bytes/Transfer
                 Time in kbytes/sec
                 DNS_SETUP_TIME - Metric 1 - Time to resolve
                 hostname through DNS
                 TRANSFER_TIME - Metric 4 - Overall time it
                 took for the data transfer
                 only
                 TRANS_BYTES - Metric 5 - The number of bytes
                 transferred

While this invention has been described in terms of several embodiments, there are alterations, permutations, and equivalents which fall within the scope of this invention. It should also be noted that there are many alternative ways of implementing the methods and apparatuses of the present invention. It is therefore intended that the following appended claims be interpreted as including all such alterations, permutations, and equivalents as fall within the true spirit and scope of the present invention.

Claims

1. A method for generating alerts from a periodically sampled system, comprising: maintaining a sampling window having a predefined number of most-recently collected samples;calculating an alert value, said alert value representing a ratio of a first value to a total number of samples in said sampling window, said first value representing a total number of samples in said sampling window having parameters exceeding a predefined sample alert threshold; andgenerating a first alert if said alert value exceeds a predefined alert threshold, wherein said first alert is generated for a sampling period that results in said alert value exceeding said predefined alert threshold only if an alert is not generated in an immediately preceding sampling period.

2. The method of claim 1 wherein said first value further includes a total number of samples in said sampling window having parameters that are equal to said predefined sample alert threshold.

3. The method of claim 1 wherein said maintaining said sampling window includes adding a most-recently collected sample to said sampling window after said most-recently collected sample is obtained and removing an oldest sample from said sampling window responsive to one of said adding said most-recently collected sample and anticipating adding said most-recently collected sample.

4. The method of claim 1 further including generating a second alert if a first condition and a second condition are met, said first condition being characterized by a generation of said first alert in a first sampling period, said second condition being characterized by an alert value X calculated with a sample obtained in a second sampling period taken into account, said second sampling period being immediately succeeding said first sampling period, said alert value X being lower than said predefined alert threshold.

5. The method of claim 1 wherein said first alert reflects a most severe violation encountered among samples of said sampling window.

6. The method of claim 1 wherein said predefined sample alert threshold comprises a plurality of threshold values.

7. The method of claim 6 further including generating a second alert if a first condition and a second condition are met, said first condition being characterized by a generation of said first alert in a first sampling period, said second condition being characterized by an alert value X calculated with a sample obtained in a second sampling period taken into account, said second sampling period being immediately succeeding said first sampling period, said alert value X being higher than said predefined alert threshold and associated with a second threshold value of said plurality of threshold values, said second threshold value being different from a first threshold value of said plurality of threshold values that is associated with said first alert.

8. The method of claim 7 wherein said second alert is generated only if said second threshold value reflects a higher level of severity than a level of severity associated with said first threshold value.

9. An article of manufacture comprising a program storage medium having computer readable code embodied therein, said computer readable code being configured to generate alerts from a periodically sampled system, comprising: computer readable code for maintaining a sampling window having a predefined number of most-recently collected samples;computer readable code for calculating an alert value, said alert value representing a ratio of a first value to a total number of samples in said sampling window, said first value representing a total number of samples in said sampling window having parameters exceeding a predefined sample alert threshold; andcomputer readable code for generating a first alert if said alert value exceeds a predefined alert threshold, wherein said first alert is generated for a sampling period that results in said alert value exceeding said predefined alert threshold only if an alert is not generated in an immediately preceding sampling period.

10. The article of manufacture of claim 9 wherein said first value further includes a total number of samples in said sampling window having parameters that are equal to said predefined sample alert threshold.

11. The article of manufacture of claim 9 wherein said computer readable code for maintaining said sampling window includes computer readable code for adding a most-recently collected sample to said sampling window after said most-recently collected sample is obtained and computer readable code for removing an oldest sample from said sampling window responsive to one of said adding said most-recently collected sample and anticipating adding said most-recently collected sample.

12. The article of manufacture of claim 9 further including computer readable code for enabling generating a second alert if a first condition and a second condition are met, said first condition being characterized by a generation of said first alert in a first sampling period, said second condition being characterized by an alert value X calculated with a sample obtained in a second sampling period taken into account, said second sampling period being immediately succeeding said first sampling period, said alert value X being lower than said predefined alert threshold.

13. The article of manufacture of claim 9 wherein said first alert reflects a most severe violation encountered among samples of said sampling window.

14. The article of manufacture of claim 9 wherein said predefined sample alert threshold comprises a plurality of threshold values.

15. The article of manufacture of claim 14 further including computer readable code for enabling generating a second alert if a first condition and a second condition are met, said first condition being characterized by a generation of said first alert in a first sampling period, said second condition being characterized by an alert value X calculated with a sample obtained in a second sampling period taken into account, said second sampling period being immediately succeeding said first sampling period, said alert value X being higher than said predefined alert threshold and associated with a second threshold value of said plurality of threshold values, said second threshold value being different from a first threshold value of said plurality of threshold values that is associated with said first alert.

16. The article of manufacture of claim 15 wherein said second alert is generated only if said second threshold value reflects a higher level of severity than a level of severity associated with said first threshold value.

17. A method for fine-tuning an alert signal obtained from a periodically sampled system, comprising: providing an alert signal generator, said alert signal generator having first logic for maintaining a sampling window at a constant window size, said sampling window containing most-recently collected samples,second logic for calculating an alert value, said alert value representing a ratio of a first value to a total number of samples in said sampling window, said first value representing a total number of samples in said sampling window having parameters exceeding a sample alert threshold, andthird logic for generating a first alert if said alert value exceeds an alert threshold, wherein said first alert is generated for a sampling period that results in said alert value exceeding said alert threshold only if an alert is not generated in an immediately preceding sampling period; andproviding a control panel having a plurality of user-manipulable controls, said plurality of user-manipulable controls including at least one of a first control for varying said window size, a second control for varying said sample alert threshold, and a third control for varying said alert threshold.

18. The method of claim 17 wherein said first alert reflects a most severe violation encountered among samples of said sampling window.

19. The method of claim 17 wherein said periodically sampled system represents an internet service monitoring system.