Java Agent v2.2.0

Release date: 2022-11-01

New features

Added the function to generate the logsink data in the virtual application.
Added the function to connect the stdout log of the logsink with txid information.
Added the data collection function for 2 seconds.
whatap.conf
```
fast_perf_enabled=false
fast_perf_interval=2000
fast_perf_gc_enabled=true
fast_perf_os_enabled=true
debug_fast_perf_enabled=false
```
Collection data
- Elapsed time
- Active Transaction
- gc information: number of gc cases, gc elapsed time, number of old generation gc cases, old generation gc elapsed time
- OS info

Added the function to send notifications based on the number of active transactions.

whatap.conf
# Whether the active transaction alerts have been set
too_many_actx_alert_enabled=false

# Ignorance time for active transaction alerts (ms)
too_many_actx_alert_silent_time=30000

# Number of active transaction alerts
too_many_actx_alert_count=50

# Number of displayed active transaction alert messages
too_many_actx_alert_msg_url_count=5

# Time for active transaction alerts
too_many_actx_alert_over_time=30000

The number of active transactions is determined by too_many_actx_alert_count = (too_many_actx_alert_over_time / 5000).
When the number of active transactions equals to or exceeds the threshold, notifications are sent by adding URL as many as too_many_actx_alert_msg_url_count to the alert message.

Profile Debug option: Added the function to check the stack in the profile and agent log when a debug target pattern is registered.
whatap.conf
```
hook_method_debug_enabled=true
hook_method_debug_patterns=
```
Adding the plug-in using class
- Added the HashTable class to be used in the {AGENT_HOME}/plugin/{...}.x file.
- If whatap.agent.api.Ref.h is used, HashTable can be used.
  Example
```
Class hashTable = (Class) whatap.agent.api.Ref.h;
```

Adding the dedicated classes for TTA BMT: Can be used to demonstrate the Java agent.

Java Class
# map
bmt.m.MyMap

# datasource
bmt.ds.MyDataSource
bmt.ds.MyConnection

# jdbc
bmt.jdbc.Config
bmt.jdbc.FakeBlob
bmt.jdbc.FakeClob
bmt.jdbc.FakeConnection
bmt.jdbc.FakeDatabaseMetaData
bmt.jdbc.FakeDriver
bmt.jdbc.FakeCallableStatement
bmt.jdbc.FakePreparedStatement
bmt.jdbc.FakeResultSet
bmt.jdbc.FakeStatement
bmt.jdbc.FakeSocket
bmt.jdbc.Lock

Tracing
// Connection Pool tracing
// ConnectionPoolASM
target.put("bmt/ds/MyDataSource", MyDataSource.className);

// Connection Pool tracing
// DBConCountCollectorThread
add(new MyDataSource());

// JDBC Connection tracing
// JDBCConnectionOpenASM
AsmUtil.add(reserved, "bmt/ds/MyDataSource", "getConnection()Ljava/sql/Connection;");

// JDBC Statement tracing
// JDBCStatementASM
target.add("bmt/jdbc/FakeStatement");

// JDBC PreparedStatement tracing
// JDBCPreparedStatementASM 
target.add("bmt/jdbc/FakePreparedStatement");

// JDBC ResultSet tracing
// JDBCResultSetASM
target.add("bmt/jdbc/FakeResultSet");

Error step
- Added the function to add the error step to the profile.
- Added the step to display error-related data in the profile by adding settings.
whatap.conf
```
profile_error_step_enabled=false
```

Adding the JMX setting: Added the function to additionally view the JMX data.

whatap.conf
perfx_tomcat_enabled=false
perfx_jmx_value_limit=100
perfx_jmx_tag_from_objectname_enabled=false

By collecting the JMX data of tomcat, the settings for checking with test1 and test2 in metrics lookup are as follows:

whatap.conf
perfx_jmx_enabled=true
perfx_jmx@test1=Catalina:type=Connector,*
perfx_jmx@test2=java.lang:type=GarbageCollector,*
perfx_jmx_ignore@test2=LastGcInfo
# perfx_debug_enabled=true
# perfx_jmx_tag_from_objectname_enabled=false

Using the OSHI library, added the function to trace the proc_mem_rss data of the process.
whatap.conf
```
oshi_proc_mem_enabled=false
```
Added the JDBC statement tracing function in org.edb.jdbc.PgCallableStatement.
Function to trace SQL parameters in the Oracle procedure
- Added the function to trace SQL parameters in Oracle procedures.
- Added the function to trace SQL parameters in CallableStatement.
In case of calling HTTP from Weblogic to HttpURLConnection, added the function to trace weblogic.net.http.HttpURLConnection.
whatap.conf
```
HttpURLConnection_weblogic=true
```
Added the function to collect the native memory data (proc_mem_rss) of the process by using the sigar library.
Added the function to collect TPS data collected with the 30-second average, in the 5-second average.
whatap.conf
```
service_metrics_spike_enabled=false
```

Added the compressed transfer function for the profile data.

Profile data is sent when the profile data exceeds the size of profile_zip_max_buffer_size.
Profile data is sent when data is accumulated in the queue or the value of profile_zip_max_wait_time is exceeded.
When the profile data is transmitted, the data is compressed except when the profile data in the queue is less than profile_zip_min_size.

whatap.conf
# Whether the profile data is compressed
 profile_zip_enabled=true

 # Queue size that has the profile data
 # ZipProfileThread.java determines the request queue size
 profile_zip_queue_size=1000

 # Waiting time for profile data transfer
 # If 1000 ms is exceeded, the profile data is sent.
 profile_zip_max_wait_time=1000

 # Buffer size of the profile data
 # If the profile data exceeds 1 MB, data is transferred.
 profile_zip_max_buffer_size=1024*1024

 # Minimum size to determine whether or not to compress the profile data
 # If the profile data is less than 100 bytes, compression is not followed.
 profile_zip_min_size=100

 # Whether or not to debug the profile data compression
 # Outputs the data related to profile data transmission (ZipPack status, recordCount, buffer size, profile data queue size) to the agent log
 debug_profile_zip_enabled=false

 # Output interval for profile data compression debug log
 # When the PROFIAL data compression debug option is turned on, the information about profile data transmission is output to the agent log at an interval of 5000 ms. 
 debug_profile_zip_interval=5000

Change (update)

Modified the option name from max_wait_time that is the logsink data transfer cycle, to logsink_max_wait_time.
Replaced the CustomCounter function collected by the plug-in with CustomTaskLoader instead of CustomCounterTask.
whatap.conf
```
custom_task_enabled=true
custom_task_jarfile={full_path}
```
Added the function to output logs when the server time difference between the agent and proxy is 10 seconds or more.
Measuring the elapsed time
- Modified to use the JVM time rather than the system time when measuring the elapsed times for transactions and profiles.
- When measuring the elapsed time, modified to use System.nanoTime() in System.currentTimeMillis().

Modified to display the time difference between servers seen in the agent log when there is a difference in the server time between the agent and proxy.

Agent log
# In case the time difference is 30 seconds or more 
"fatal warning time-reverse (time difference) ms"

# In case the time difference is less than 30 seconds
"warning time-reverse (time difference) ms"

Modified the options for user counting in the form of wclient_xxx. The previous option names can be used but new option names are recommended, starting from v2.2.0.

whatap.conf
# Replacement
trace_user_method=IP
wclient_trace_type=IP

# Replacement
trace_user_jsession_key=JSESSIONID
wclient_jsession_key=JSESSIONID

# Replacement
trace_user_header_key=token
wclient_header_key=token

# Replacement
trace_user_cookie_limit=4000
wclient_cookie_limit=4000

# Replacement
trace_user_cookie_domain=
wclient_cookie_domain=

# Replacement
realtime_user_thinktime_max=300000
wclient_thinktime_max=300000

# Replacement
trace_user_using_ip=true
wclient_using_ip=true

# Replacement
user_header_ticket=
wclient_header_ticket=

# Replacement
wclient_max_count=70000

Recompiled the embedded weaving component by applying the count settings related to the number of users.
Modified to use System.nanoTime() when calculating the CounterPack elapsed time.

Included the release version in the log output logo.

Logo
  _       ____  ______           
| |     / / /_/_  __/___ _____   ___  ___ 
| | /| / / __ \/ / / __ `/ __ \ | _ \|_  ) 
| |/ |/ / / / / / / /_/ / /_/ / |   / / /
|__/|__/_/ /_/_/  \__,_/ .___/  |_|_\/___|
                      /_/                    
The Best Observability in Java

Modified the default value for the Java agent’s statistics collection from 5000 to 10000, checking every 2 seconds when sending statistics data.
whatap.conf
```
stat_tx_max_count=10000
```
Updated the version of the OSHI library that collects performance metrics from oshi-core-5.4.1 to oshi-core-6.1.6.
Changed the library folder name of {AGENT_HOME} that collects performance metrics.
- Renamed the SIGAR library folder: lib -> lib3.
- Renamed the OSHI library folder: lib2 -> lib4.

Fixed (bug fixing)

Fixed a bug that occurred when adding headers related to multi-transaction to the read-only HttpHeaders of the Webflux weaving plug-in, an embedded Java agent plug-in.

Applied to webflux-5.1, webflux-5.2, and webflux-5.3.
Fixed the bug that occurred when the elapsed time of transactions and steps was less than 0.
Fixed the bug where the elapsed time is displayed as 0, when the elapsed time of transactions and steps is less than 0.
Fixed the blocking_detect_count handling bug.

New features​

Change (update)​

Fixed (bug fixing)​

New features

Change (update)

Fixed (bug fixing)