Chat now with support
Chat con el soporte

Foglight Experience Monitor 5.8.1 - Metric Reference Guide

Foglight Experience Monitor metric categories
General metric categories Application-oriented categories Service-oriented categories Web services category Content-oriented categories Infrastructure-oriented categories Types of metrics How standard metrics are calculated End-User Access Speed Network Latency Instrumented metric: Load Time Time frames in which metrics are viewed
Metric definitions
Non-system category metrics
Access Speed Access Speed Distribution Alarm Application Time Cipher Suite Not Found City Clicks per Session Client IP Command Byte Volume - Server Command Byte Volume - User Command Client Time Command Completion Time Command Count Command Data Size - Server Command Data Size - User Command Initial Response Time Command Initial Response Time Distribution Command Network Latency Command Processing Time Command Processing Time Percentile Command Processing Time Service Level Command Timeout Count Component List Size Connection Closed Count Connection Count Connection Duration Connection Established Count Connection Open Wait Count Connection Reset Count Connection Timeout Count Duration End-to-End Time for Dialup Users (0-56k) End-to-End Time for ISDN Users (56k-128k) End-to-End Time for Low-End Broadband Users (128k-384k) End-to-End Time for High-End Broadband Users (384k-728k) End-to-End Time for T1/Lan Users (> 728k) End-to-End Time Service Level Error Count (HTTP 4xx Client Errors) Error Count (HTTP 5xx Server Errors) Fault Actor Fault Code Fault Detail Fault String Fault Subcodes Field Out of Range Hit Count Hit End-To-End Time Hit Redirect Ratio Hit Size Incomplete SSL Key Invalid SSL Content Invalid SSL Handshake Type Invalid SSL Key Invalid SSL Record Length Invalid SSL State Invalid SSL Version Number IP Packet Byte Volume - Server IP Packet Byte Volume - User IP Packet Count - Server IP Packet Count - User IP Packet Size - Server IP Packet Size - User ISP Load Time Load Time Percentile Login Name Missing Reusable Secret Missing SSL Key Other Requests Page Access Speed Distribution Page Client Time Page Connection Count Page Download Attempts Page Download Size Page Element Count Page End-to-End Time Page End-To-End Time Distribution Page End-To-End Time Percentile Page Exit Count Page Network Latency Page Peak Count per Second Page Processing Time Page Processing Time Distribution Page Processing Time Percentile Page Redirect Ratio Page Stop Count Page Stop Rate Page Stop Time Page Think Time Page Timeout Count Page Views Path Count Path Duration Path End-To-End Time Path End-to-End Time Percentile Processing Load Percentage Processing Time Service Level Referrer Request Code Count Response Code Count Service Abort Count Service Client Time Service Count Service Duration Service End-to-End Time Service End-to-End Time Distribution Service End-to-End Time Percentile Service Network Latency Service Peak Count per Second Service Processing Load Percentage Service Processing Time Service Processing Time Distribution Service Processing Time Percentile Service Processing Time Service Level Service Request Count per Second Service Started Count Service Think Time Service Timeout Count Session Count Session Key Site Stickiness SOAP Operation Client Time SOAP Operation Completion Time SOAP Operation Completion Time Distribution SOAP Operation Consumer Overhead Ratio SOAP Operation Count SOAP Operation End-To-End Time SOAP Operation End-to-End Time Distribution SOAP Operation End-to-End Time Percentile SOAP Operation Failure Count SOAP Operation Initial Response Time SOAP Operation Network Latency SOAP Operation Peak Count per Second SOAP Operation Processing Time SOAP Operation Processing Time Distribution SOAP Operation Processing Time Percentile SOAP Operation Provider Overhead Ratio SOAP Operation Success Count SOAP Operation Success Ratio SOAP Operation Think Time SOAP Operation Timeout Count SOAP Transaction Abort Count SOAP Transaction Client Time SOAP Transaction Completion Time SOAP Transaction Completion Time Distribution SOAP Transaction Count SOAP Transaction Duration SOAP Transaction End-to-End Time SOAP Transaction End-to-End Time Distribution SOAP Transaction End-to-End Time Percentile SOAP Transaction Initial Response Time SOAP Transaction Network Latency SOAP Transaction Peak Count per Second SOAP Transaction Processing Time SOAP Transaction Processing Time Distribution SOAP Transaction Processing Time Percentile SOAP Transaction Processing Time Service Level SOAP Transaction Request Peak Count per Second SOAP Transaction Think Time SOAP Transaction Timeout Count SSL Memory Error Step Started Count Step Completion Count Step Abort Count Step Timeout Count Subnet Success Ratio Success Ratio (with Client Errors) Success Ratio (with Server Errors) TCP Command Client Time TCP Command Completion Time TCP Command Count TCP Command Initial Response Time TCP Command Initial Response Time Percentile TCP Command Network Latency TCP Command Processing Time TCP Connection Open Peak Count per Second TCP Queue Size TCP Control Segment Byte Volume - Client TCP Control Segment Byte Volume - Server TCP Control Segment Count - Client TCP Control Segment Count - Server TCP Control Segment Size - Client TCP Control Segment Size - Server TCP Data Segment Byte Volume - Client TCP Data Segment Byte Volume - Server TCP Data Segment Count - Client TCP Data Segment Count - Server TCP Data Segment Size - Client TCP Data Segment Size - Server TCP Duplicate Resend Segment Byte Volume - Client TCP Duplicate Resend Segment Byte Volume - Server TCP Duplicate Resend Segment Count - Client TCP Duplicate Resend Segment Count - Server TCP Duplicate Resend Segment Size - Client TCP Duplicate Resend Segment Size - Server TCP Fragmented Segment Byte Volume - Client TCP Fragmented Segment Byte Volume - Server TCP Fragmented Segment Size - Client TCP Fragmented Segment Size - Server TCP Fragmented Segment Count - Client TCP Fragmented Segment Count - Server TCP Invalid Checksum Segment Byte Volume - Client TCP Invalid Checksum Segment Byte Volume - Server TCP Invalid Checksum Segment Size - Client TCP Invalid Checksum Segment Size - Server TCP Invalid Checksum Segment Count - Client TCP Invalid Checksum Segment Count - Server TCP Payload Byte Volume - Client TCP Payload Byte Volume - Server TCP Resend Segment Byte Volume - Client TCP Resend Segment Byte Volume - Server TCP Resend Segment Count - Client TCP Resend Segment Count - Server TCP Resend Segment Size - Client TCP Resend Segment Size - Server TCP Segment Byte Volume - Client TCP Segment Byte Volume - Server TCP Segment Count - Client TCP Segment Count - Server TCP Segment Size - Client TCP Segment Size - Server TCP Timeout Count - Syn TCP Timeout Count - Idle TCP Timeout Count - Half Open TCP Urgent Segment Byte Volume - Client TCP Urgent Segment Byte Volume - Server TCP Urgent Segment Count - Client TCP Urgent Segment Count - Server TCP Urgent Segment Size - Client TCP Urgent Segment Size - Server Unexpected Client Hello Message Unexpected Server Hello Message Unknown Key Exchange Unload Time Unload Time Percentile User Agent User Count User Stickiness
System category metrics
Agent Memory Consumption Warnings Agent Restarts Baseline Partition Size Backup Errors Checksum Errors Client Segments Count Client Segments Missing Client Segments Missing Rate Communication Errors Connection Errors Connections Started Connections Dropped CPU1 Temperature CPU2 Temperature CPU Core 1 Utilization CPU Core 2 Utilization CPU Core 3 Utilization CPU Core 4 Utilization CPU Core 5 Utilization CPU Core 6 Utilization CPU Core 7 Utilization CPU Core 8 Utilization CPU Utilization Database Load Errors Database Load Time Database Partition Usage Database Records Loaded Datastore Inode Usage Datastore Partition Usage Day Partition Size Excessive Data Volume FxV Bytes Transmitted FxV Discards FxV Hits Transmitted Hour Partition Size Log Partition Usage Memory Consumption Warnings Memory Utilization Metric Table Entries Metric Table Size Minute Partition Size Monthly Partition Size OS Partition Usage Packet Drop Rate Packet Errors Packet Errors - Port 1 Packet Errors - Port 2 Packet Errors - Port 3 Packet Errors - Port 4 Packet Errors - Port 5 Packet Errors - Port 6 Packet Errors - Port 7 Packet Errors - Port 8 Packets Captured Packets Captured – Port 1 Packets Captured – Port 2 Packets Captured – Port 3 Packets Captured – Port 4 Packets Captured – Port 5 Packets Captured – Port 6 Packets Captured – Port 7 Packets Captured – Port 8 Packets Dropped Packets Dropped – Port 1 Packets Dropped – Port 2 Packets Dropped – Port 3 Packets Dropped – Port 4 Packets Dropped – Port 5 Packets Dropped – Port 6 Packets Dropped – Port 7 Packets Dropped – Port 8 Packets Processed Packets Processed – Port 1 Packets Processed – Port 2 Packets Processed – Port 3 Packets Processed – Port 4 Packets Processed – Port 5 Packets Processed – Port 6 Packets Processed – Port 7 Packets Processed – Port 8 RAID Array Degraded Relation Table Entries Relation Table Size Resource Table Entries Resource Table Size SiteMinder Errors Server Segments Count Server Segments Missing Server Segments Missing Rate SOAP Error Rate SOAP Errors SOAP Message Count SOAP XML Errors SSL Connection Errors SSL Connections Released SSL Connections Started SSL Connection Error Rate Swap Partition Usage System Temperature Tmp Partition Usage Total Table Size Value Count Table Entries Value Count Table Size Weekly Partition Size

Calculating access speed from the client side

The following diagram shows the standard method to calculate the user’s access speed, assuming there is client instrumentation to measure the time it takes to receive the server packets.

In this diagram, the client begins receiving a data block and starts a timer. When the client receives the last byte of the data block, the timer is stopped. The total number of bits received by the client during that period is divided by the elapsed time to determine access speed.

Calculating access speed from the server side

Access speed from a passive monitor on the server side of the connection is calculated differently than the standard client instrumentation method outlined in the previous section.

Calculating access speed from the server side means overcoming several challenges:

To deal with these challenges, access speed can be calculated by using characteristics of the TCP protocol, as well as the Round-Trip Time metric.

Whenever the server sends two maximum-sized packets, the client must immediately acknowledge it has received them. This is a requirement of the TCP protocol called delayed acknowledgements. The following diagram illustrates this case.

The following diagram illustrates the issue of how a network delay inflates the calculation of access speed.

To overcome this problem, the system can use the Round-Trip Time metric (as described in the next section). Thus, the access speed calculation is as follows:

To improve the accuracy of the End User Access Speed metric, the above calculation is performed repeatedly throughout the user’s session whenever two valid maximum-sized TCP segments are transmitted. The Round-Trip Time metric is also kept current. This repeated sampling means that at the end of the user session:

Network Latency

How long does it take a packet to traverse the network from the user to the web site and from the web site back to the user?

Network latency measures one way network delay, and is calculated by equal halves of the following: the amount of time for packets to be sent, and the amount of time it takes for the TCP acknowledgement of the packet to be received. The time it takes to send data and receive an acknowledgment (ACK) is often called the round-trip time. To measure round-trip times, network administrators often utilize a utility named ping that transmits a packet to a specific IP address, and then measures the amount of time required for the acknowledgement packet to be received.

As a passive monitoring solution, the appliance does not transmit packets on the network to calculate network latency. Instead, it uses a characteristic of the TCP protocol stack to accomplish this.

In a TCP connection, each packet that contains data must be acknowledged by means of another packet (the Acknowledgement packet or ACK) from the receiving side. Another useful characteristic of the TCP protocol is that the receiving side is required to immediately send an acknowledgement for every two full data packets it receives. This sending of acknowledgements is how TCP controls data congestion.

By measuring the delay from when a data packet is sent until the acknowledged is received, a passive monitor can calculate the network latency:

For data packet (i):

Since the monitoring appliance is always positioned on the server side of the conversation, the appliance looks at data packets sent by the server and then times how long until an acknowledgement is received from the client for that data. Since round-trip times can vary throughout the life of a TCP connection, the samples for a round-tip time are averaged into a smoothed round-trip time estimate (SRTT).

These data packets that are traveling back and forth on the TCP connection form a command. When a command is completed, the current smoothed round-trip time is used to set the Command Network Latency:

In the web console, on the Configure > URLs > Advanced Options page, the “Ignore delayed ACKs in network latency metric calculations” option causes the system to exclude delayed ACKs (TCP acknowledgements) from the calculations for this metric. This has the benefit of providing more accurate latency metrics, but can also cause the value for this metric to sometimes be reported as zero. WIth this option enabled, only immediate client ACKs are used in the calculations for this metric. These immediate ACKs are only mandatory if the client receives more than a single data packet from the server. If the message payload fits entirely into a single packet, the client is not required to send the immediate ACK. Consequently, the value reported for this metric can be zero if there were no immediate ACKs observed within the reporting time period.

In the web console, on the Configure > URLs > Advanced Options page, the “Use unweighted samples for network latency metric calculation” option causes the system to use an unweighted formula for this metric calculation. The unweighted formula is essentially a running total of all latency measurements that have occurred so far in the TCP connection, divided by the number of measurements. This is a direct average of all latency measurements in the TCP connection, unlike the “smoothed” latency measurement described in the Calculation section.

More related topics:

How Page and Command Network Latency differ

Both Page and Command Network Latency are both derived from the underlying round-trip time estimates. However, how those round trip times are combined does differ.

Command Network Latency is derived from the smoothed round-trip time for the single TCP connection that the command was transmitted on (as described in the previous section).

A page can consist of multiple commands sent across multiple TCP connections. In this case the Page Network Latency equals the average of the network latency of its component commands.

Documentos relacionados