499-01 Riverbed Certified Solutions Professional – Application Performance Management Questions and Answers

AppResponse Xpert

Browser plugin

A load balancer such as Stingray

SysAdmin or webmaster

Proxy server

Use Riverbed Steelhead to inject the script at run time from the web server.

Use Riverbed Stingray to inject the script at run time from the web server.

Manually paste the Javascript in any web page you want to monitor with BrowserMetrix.

Edit a master file for the application by adding the Javascript to the header section of the file.

By having many people log in with the same account

By clicking on the Link tool to create a unique URL for the dashboard

By exporting the dashboard as a report

AppResponse Xpert

AppInternals Xpert

Cascade Profiler

BrowserMetrix

Select all of the desired metrics at panel creation time.

Click on the "+" icon on the panel after you have chosen an initial set of metrics.

Drag one panel on top of another to merge them.

AppSensor Xpert, AppInternals Xpert, and BrowserMetrix

AppTransaction Xpert, AppInternals Xpert, and BrowserMetrix

AppResponse Xpert, AppInternals Xpert, and BrowserMetrix

AppSensor Xpert, AppResponse Xpert, and BrowserMetrix

After you select a host in the search field

From the sidebar only

By right-clicking on the main page

By right-clicking a tab

A change in direction in the data exchange

One full TCP data request-response cycle

A TCP connection that finishes with a FIN-ACK connection close

The TCP three-way handshake (SYN, SYN-ACK, ACK)

Data sent between TCP keepalives

Database server name or internet protocol (IP) address

Database type

Port number(s)

Table name

Query type

It is based on packet loss reported on routers.

It is useful to correlate packet loss with round trip time (RTT).

I should not check against packet retransmission rate.

It is based on the percentage of packets from transmission control protocol (TCP) servers that were retransmitted.

Inbound traffic to a group and outbound traffic from a group.

The number of turns between servers and clients.

Average size of data transfers to and from business groups.

AppResponse Xpert appliances do not report on utilization.

number of TCP connections / min (for HTTP/HTTPS sessions only)

number of total page views / min

number of HTTP requests / min

number of unique URLs / min

Large network transfer effects and large protocol delays

A high number of application turns and application messages

A high percentage of tier processing effects delay

Large parallel effects delay

Network effects

Tier processing effects

User delay

TCP/IP delay

CPU utilization and memory leaks

Network, packet loss and parallel effects

Network, latency and server

The default pie chart displays only information about flows

Giant packets, runts and corrupted packets

Red

Green

Blue

Yellow

Purple

This transaction takes about 4 seconds to complete and this analysis proves the user's belief that additional WAN bandwidth would improve application performance.

The majority of activity is between the two servers.

Most of the traffic consists of small messages.

Much of the application response time is caused by the Wide Area Network (WAN).

This is not feasible currently.

Recompile or modify the code.

Set the JAVA_TOOL_OPTIONS environment variable to add the argument into the application start up.

We should not instrument applications that use embedded virtual machines.

10

20

50

100

200

MovingWindowPercentile

MovingWindowSigma

HourlyLimits

HistoricalBaseline

HardLimits

Some code compilation or modification is needed.

You can set the IBM_JAVA_OPTIONS environment variable to add the argument into the application startup.

IBM SDK Java is not supported.

Setting the IBM_JAVA_OPTIONS environment variable will add the same argument into every IBM SDK Java application startup.

Symmetric multiprocessor (SMP) managed nodes on the network

Virtual machines specified in the Trace Transaction Warehouse interface

Managed nodes specified by SMP

Any server that is network accessible

The Web tier is the bottleneck because it has the most open requests waiting for a response.

The DB tier is the bottleneck because it is the lowest tier that is clearly not handling requests in time.

The App tier is the bottleneck because it is the first tier not handling the same number of requests as the tier above it.

The Web tier is the bottleneck because at an Offered Load of 300, there are still 50 requests waiting from the Web tier, and it never recovers. The lower tiers do not get the requests passed to them because the Web tier is not processing fast enough.

The App tier is the bottleneck because the DB tier is handling everything the App tier sends it.

End-user experience monitoring

User-defined transaction profiling

Application component deep dive

Network device monitoring

Status polling

Use the built-in email and SNMP mechanism or a custom forensic script in AppInternals Xpert to send alerts to the existing system.

Configure alerts in AppResponse Xpert to send SNMP traps to the existing system.

Set up BrowserMetrix to send email alerts to an email gateway.

Add email alerts to the RPM Dashboard settings.

Configure Transaction Trace Warehouse to send email alerts to an email gateway when certain thresholds are met.

AppInternals Xpert

AppResponse Xpert

AppTransaction Xpert

AppSensor Xpert