Configure ColdFusion Performance Monitoring Toolset settings

General

Alerts

The mail settings here applies to all nodes, clusters, and groups for which monitoring is enabled, and the mail settings, which are not configured in Alerts.

However, if you want to change mail notification settings for a particular group, node, or cluster, then click Alerts > Settings, and change the values accordingly.

Datastore settings

ColdFusion servers

Untracked servers

These are the servers that were not being tracked by the toolset when you started the toolset for the first time.

You can add the servers manually to be discovered by the toolset.

Toggle the Auto-discovery switch to discover any new ColdFusion servers added in the network.

Tracked servers

These are the servers that were automatically discovered by the toolset when starting up.

You can:

• Search for a tracked server
• Send a server back to the untracked servers list
• See a list of other servers that are tracked by another instance of Performance Monitoring Toolset.

Groups

There are situations when you only want to monitor a handful of sites. Instead of monitoring them separately, you can combine them and create a group. You can then apply the monitoring settings on the new group.

You can modify a group and delete a group.

After deleting a group, all instances in the group (stand-alone or J2EE) reverts to being stand-alone nodes.

You can add as many instances as you can in a group.

If a site is served by two nodes, you can include the sites in a group, and then you can see how the sites are performing.

Similarly, you can include as many nodes you require in a group and monitor them collectively.

Users

Archiving

As data grows in time, the need to properly archive it grows as well.

Archiving helps you to maintain regular backups of your data and comply with an organization's data retention policies. Using archiving, you can remove obsolete production data via a well-defined process. In Performance Monitoring Toolset, you can archive records in a repository and also set the frequency of archival and data retention period.

Datastore

Delete data from Datastore

Here you can see all data based on indices, which is data grouped by date. In an index, a full day’s data is stored. You can also delete an index to free up disk space.

You can delete the request indices for the Datastore.

Typically, you’d delete an index to free up system space.

You can also delete a specific index from the list.

To delete any index in the table, click Delete.

Delete data on a schedule

You can use the scheduler to delete data periodically. The depends on the frequency and data older than options, and based on the values defined, the scheduler will run and will delete data, which is older than the specified amount of days.

The scheduler, in this example, executes every day and deletes data that is older than seven days.

By default, the value for data deletion scheduler is 7 and 1 days. The data deletion scheduler is disabled by default.

After you click Submit, the scheduler run details get displayed.

To stop the scheduler, clear the option Enable. Then, click Submit.

Updates

You can verify and make product updates from the browser-interface itself.

Verify if there are any product updates by clicking update in the left navigation panel of the Administrator console. The updates can include hotfixes and security hotfixes for ColdFusion Performance Monitoring Toolset.

Available Updates

Click Check Updates to see if any hotfix updates are available for installation.

You can either:

• Download: Downloads and places the file in <pmt_home>/hf-updates/ for installation later.
• Download and Install: Downloads the hotfix and performs a silent installation.

Installed updates

Lists all hotfix updates to Performance Monitoring Toolset that you have installed.

Settings

Provides options to specify update preferences such as update notifications or if to automatically check for updates.

If you have set up a local update site, you can also specify URL of that site to get updates.

Server/Node/Cluster Settings

Monitoring

Health score & baseline

Performance Monitoring Toolset can measure how your system is performing. In various pages of our brand new toolset, you can find health scores for node/application/cluster/group. These health scores reflect the state of various components in your ColdFusion setup.

This is how the health score is calculated.

Node Health Score

Health Score depends on four parameters. They are - ART, Error rate, CPU usage, and Memory Usage. The toolset calculates scores for each of the four parameters as shown below.

• ART Score: To calculate an ART score, the ART of the last five minutes is calculated and compared against the ART provided by you.
• For example, if A_baseline  is the baseline ART provided you and A_actual is ART of last five minutes and let (2*A_baseline  - A_baseline) / 5 = diff, then,
• • ART Score = 100    if A_{actual  <=}  A_baseline
  • ART Score = 83      if A_baseline < A_actual  <=   A_baseline + diff
  • ART Score = 66      if A_baseline + diff    < A_actual  <=   A_baseline + 2*diff
  • ART Score = 50      if A_baseline + 2*diff < A_actual  <=   A_baseline + 3*diff
  • ART Score = 33      if A_baseline + 3*diff < A_actual  <=   A_baseline + 4*diff
  • ART Score = 16      if A_baseline + 4*diff < A_actual  <=   A_baseline + 5*diff
  • ART Score = 0        if A_{actual  >}   A_baseline + 5*diff
• Error Score : To calculate Error score, the Error percentage of last five minutes is calculated and compared against error percentage provided by the user.
• For example, if E_baseline  is error baseline provided by user and E_actual is error percentage of last 5 minutes and let (Min(5*E_baseline  , 100) -  E_baseline) / 5 = diff, then,
• • Error Score = 100    if E_{actual  <=}  E_baseline
  • Error Score = 83      if E_baseline < E_actual  <=   E_baseline + diff
  • Error Score = 66      if E_baseline + diff  < E_actual  <=   E_baseline + 2*diff
  • Errror Score = 50    if E_baseline + 2*diff < E_actual  <=  E_baseline + 3*diff
  • Error Score = 33      if E_baseline + 3*diff < E_actual  <=   E_baseline + 4*diff
  • Error Score = 16      if E_baseline + 4*diff < E_actual  <=   E_baseline + 5*diff
  • Error Score = 0        if E_{actual  >}  E_baseline + 5*diff
• CPU Score : To calculate the CPU score, CPU usage percentage of last 5 minutes is calculated and compared against cpu usage percentage provided by the user.
• For example, if C_baseline  is the CPU usage baseline provided by you and C_actual is the CPU usage percentage of last 5 minutes, and let (Min(5*C_baseline  , 100) -  C_baseline) / 5 = diff, then,
• • CPU Score = 100    if C_{actual  <=}  C_baseline
  • CPU Score = 83      if C_baseline < C_actual  <=   C_baseline + diff
  • CPU Score = 66      if C_baseline + diff  < C_actual  <=  C_baseline + 2*diff
  • CPU Score = 50      if C_baseline + 2*diff < C_actual  <=  C_baseline + 3*diff
  • CPU  Score = 33     if C_baseline + 3*diff < C_actual  <=  C_baseline + 4*diff
  • CPU Score = 16      if C_baseline + 4*diff < C_actual  <= C_baseline + 5*diff
  • CPU Score = 0        if C_{actual  >}  C_baseline + 5*diff
• Memory Score :  To calculate the memory score, the heap usage percentage of the last five minutes is calculated and compared against heap usage percentage provided by the user.
• For example, if M_baseline  is heap usage baseline provided by user and M_actual is heap usage percentage of last 5 minutes and let (Min(5*M_baseline  , 100) -  M_baseline) / 5 = diff, then,
• • Memory Score = 100    if M_{actual  <=}  M_baseline
  • Memory Score = 83      if M_baseline < M_actual  <=   M_baseline + diff
  • Memory Score = 66      if M_baseline + diff  < M_actual  <=  M_baseline + 2*diff
  • Memory Score = 50      if M_baseline + 2*diff < M_actual  <=  M_baseline + 3*diff
  • Memory Score = 33       if M_baseline + 3*diff < M_actual  <=  M_baseline + 4*diff
  • Memory Score = 16      if M_baseline + 4*diff < M_actual  <= M_baseline + 5*diff
  • Memory Score = 0        if M_{actual  >}  M_baseline + 5*diff
• Node Health Score = (ART Score * ART Weightage + Error Score * Error Weightage + CPU Score * CPU Weightage + Memory Score * Memory Weightage) / (ART Weightage + Error Weightage + CPU Weightage + Memory Weightage)

Note that the error score has veto power. If the error score is zero, the entire health score becomes zero.

Application Health Score

Health Score depends on two parameters. They are - ART and Error rate. The toolset calculates scores for each of the two parameters as shown below.

• ART Score : To calculate ART score, ART of last five minutes is calculated and compared against the ART provided by the user.
• Suppose, A_baseline  is ART baseline provided by user and A_actual is ART of last 5 minutes and let (2*A_baseline  - A_baseline) / 5 = diff
• • ART Score = 100    if A_{actual  <=}  A_baseline
  • ART Score = 83      if A_baseline < A_actual  <=   A_baseline + diff
  • ART Score = 66      if A_baseline + diff < A_actual  <=   A_baseline + 2*diff
  • ART Score = 50      if A_baseline + 2*diff < A_actual  <=   A_baseline + 3*diff
  • ART Score = 33      if A_baseline + 3*diff < A_actual  <=   A_baseline + 4*diff
  • ART Score = 16      if A_baseline + 4*diff < A_actual  <=   A_baseline + 5*diff
  • ART Score = 0        if A_{actual  >}  A_baseline + 5*diff
• Error Score : To calculate Error score, Error percentage of last five minutes is calculated and compared against error percentage provided by the user.
• Suppose, E_baseline  is error baseline provided by user and E_actual is error percentage of last five minutes and let (Min(5*E_baseline  , 100) -  E_baseline) / 5 = diff
• • Error Score = 100    if E_{actual  <=}  E_baseline
  • Error Score = 83      if E_baseline < E_actual  <=   E_baseline + diff
  • Error Score = 66      if E_baseline + diff  < E_actual  <=   E_baseline + 2*diff
  • Errror Score = 50    if E_baseline + 2*diff < E_actual  <=   E_baseline + 3*diff
  • Error Score = 33      if E_baseline + 3*diff < E_actual  <=   E_baseline + 4*diff
  • Error Score = 16      if E_baseline + 4*diff < E_actual  <=   E_baseline + 5*diff
  • Error Score = 0        if E_{actual  >}  E_baseline + 5*diff
• Application Health Score = (ART Score * ART Weightage + Error Score * Error Weightage) / (ART Weightage + Error Weightage)

Note that error score has veto power. If the error score is zero, the entire health score becomes zero.

Cluster Health Score :  Cluster health score is simple mean of all nodes in cluster.

Group Health Score :  Group health score is simple mean of all nodes in group.

Setting Baseline for different parameters:  In Health Score and Baseline section of settings, admin can set baseline and weightage parameters for calculation of various health scores.

For a cluster of nodes

For application

Alerts

The Alert options let you specify the thresholds for when to generate an alert. Alerts provide warnings of potential problems, including a slow server or an unresponsive server. The unresponsive-server alert is triggered when the server's average response time exceeds a specified limit. The unresponsive-server alert is triggered when more than a specified number of threads are busy for more than a specified number of seconds. The unresponsive-server alert creates a snapshot file, which lets you determine where request threads are unresponsive.

Both types of alert let you run a custom CFC when the alert is triggered. The CFC also lets you provide your own automated response to an alert condition. You can specify whether to send an e-mail notification when an alert is triggered, and to whom.

The CFC triggers the alerts. The CFC contains the functions:

• onAlertStart: This function executes when an alert becomes active. The structure passed to this function contains information about settings at which the alert was activated.
• onAlertEnd : This function executes when the server recovers from an alert or when the alert invalidates. The structure passed to this function contains the alert settings when the alert was disabled or recovered.

Delete alerts

You can delete all alerts by clicking the notification icon on the top menu. Here are the ways using which you can delete alerts.

• You can delete any specific alert by choosing the alert and clicking the Delete button.
• You can select up to 100 alerts at a time and then delete them altogether.
• You can delete all the alerts without selecting any alert. Simply click the Delete button.

On the list of all alerts, to delete any alert, click Delete near each alert.

After you click Delete, there is a confirmation pop-up message, where you click OK to delete the alert.

JVM alerts

Because ColdFusion is an enterprise Java application, the Java Virtual Machine (JVM) is the software component that most influences performance. Different JVMs from different vendors and different versions of the same JVM from the same vendor have different performance characteristics. You can benefit from changing the JVM that you are using with ColdFusion.

To configure ColdFusion to use a different JVM, edit the cf_root/runtime/lib/ jvm .config file with a text editor. To point to the root directory of the JVM to use, modify the value of java.home . Alternatively, you can switch to a different JVM in the ColdFusion Administrator on the Java and JVM Settings page.

Because switching the JVM changes the software environment significantly, do so first in a development or testing environment.

The JVM performs memory management and can have a significant effect on your performance depending on how you configure the JVM. The most important settings for the JVM are the initial heap size and maximum heap size. The initial heap size represents the amount of memory that the JVM uses on startup; the maximum heap size represents the amount of memory that the JVM can use. You can modify these settings in the ColdFusion Administrator on the Java and JVM Settings page. The Initial Memory Size setting specifies the initial heap size; the Maximum Memory Size setting specifies the maximum heap size.

The JVM arguments for initial heap size and maximum heap size are -XmsNm and -XmxNm respectively. N is the size of the heap in megabytes (MB). The JVM arguments are stored in the jvm.config file, in the value of the java.args setting.

In the Alerts section, you can tune JVM settings for the following:

• JVM CPU usage.
• JVM memory old generation.
• JVM memory metaspace.
• JVM memory code cache.
• JVM heap usage.

System alerts

Slow server

If the average response time of the server is greater than the value specified, a slow server gets created.

Timeout

If the number of requests specified by the timeout count on an average or at all time in a specific interval, a timeout is created.

Unresponsive server

If the number of threads executes for longer than the amount of time specified below, the server becomes unresponsive.

System CPU usage

Define the trigger when CPU usage more than the defined threshold within an interval.

System memory usage

If the memory usage exceeds the defined threshold, define an action to be taken.

Others

Connector

Define the trigger if the number of busy connections are more than the specified amount of total number of connections.

Settings

• Verify connection: Click this button to verify the connection details that you'd provided for the mail settings and Slack webhook.
• Clear: Clear the form and re-enter all the details.

Note:

1. Set JVM flag, -Dmailserver.security=ssl/ tsl in JVM config in Performance Monitoring Toolset.
2. Add SSL certificate in JVM in Performance Monitoring Toolset.

Restart Performance Monitoring Toolset.

Purge

Lists all active ColdFusion instances that are a part of a cluster. You can also see the following:

• Sites that are a part of the cluster.
• Applications that are a part of the cluster.

On the Server tab, if you click the Delete icon on a cluster, all instances within that cluster gets deleted. The process is irreversible. The monitoring for that cluster also stops.

Similarly,

On the Connector tab, you can delete the sites and monitoring is also stopped.

On the Applications tab, if you delete the applications, the applications cannot be rolled back for further monitoring.