The goal of this study is to create a web monitoring application that can retrieve and save data from java application server resources. This study's research methodology consists of two phases: data collecting and software development. The data gathering phase includes a review of the literature, a survey, the distribution of questionnaires, and an analysis of similar apps. The waterfall method is used during the development phase. The ultimate result of this research is a web application that can monitor resource utilization on the java application server and send email notifications if it becomes excessive. Finally, web application server monitoring can be used to make the developer's job easier.
Research
The Development of Monitoring Java Web Application Server Resource Usage using JavaMelody Promotheus Grafana for Sever Maintenance
https://doi.org/10.21203/rs.3.rs-956277/v1
This work is licensed under a CC BY 4.0 License
Version 1
posted
You are reading this latest preprint version
Web Application
Monitoring
Java
JavaMelody
Prometheus
Grafana
The rise of information technology and the internet have transformed how individuals obtain information, engage in entertainment, conduct e-commerce, and so much more. Because of such, many people claim that the present time is the beginning of the next industrial revolution 4.0. As a result of the outbreak of coronavirus (CV) in 2019 that turned into a pandemic, people rely more on the internet as a result of suddenly needing to conduct online tasks like job or school from home, or as a way to buy and sell products via the internet. A study on the total internet usage and population statistics from Nielsen Online for the year 2021 and the first quarter of 2021 It is predicted that the number of internet users around the world will be on the rise every year. World Internet Usage and Population Statistics 2021 Year-Q1 is shown below.
| Population (2021 Est) | Population % of world | Internet Users 31 Mar 2021 | Penetration Rate (% Pop.) | Growth 2000–2021 | Internet World % | |
|---|---|---|---|---|---|---|
| Total | 7.875.765.587 | 100% | 5.168.780.607 | 65.6% | 1,331.9% | 100% |
According to the Table 1, world internet growth from 2020 to 2021 is predicted approximately at 1331.9% and the penetration rate is around 65.6%. A similar pattern is seen in Indonesia, where penetration in 2020 is at 73.7% compared to 2018, where penetration was at 64.8% [2].
The worldwide rise in internet users means web server availability must be maintained. Even if the web server is used often, server availability is very significant because if it is under 90% for an entire year, it means the website is inaccessible for around 87 hours and 40 minutes, which might result in users switching to another website that can cause revenue loss. Use of web server monitoring can be used to improve and maintain server availability.
Server monitoring is a process of looking at data and may be presented in numerical or graphical form. Usage data from the web server will assist with the decision-making process of server maintenance. If we look at the overall growth of internet users and registered domains, we can safely conclude that there is an increasing need for digital products and services. Author believes to build a decision-monitoring program to support server owners. For example, here are some ways that monitoring of the server will be put to use:
This is the time to inform the web application server about its availability.
It reveals the current resource consumption of the web application server at this moment.
Allow the server owner to have all of the relevant info at their fingertips.
4. When the web application server is offline or server utilization resources exceed the limit, send alert emails to the owner.
To follow the rest of the paper, please find it organized as follows: Section 2 presents other studies that are linked to this topic, whereas section 3 provides information on system design. Section 4 showcases web server monitoring application findings. The conclusions in Section 5 are taken from data in Section 4.
Endang Ray., [3] This research aim is to provide a server monitoring solution that can be used anywhere from any mobile device. This application can be used for checking server, port, CPU, memory, and network state; as well as to detect network issues, restart hosts, services, and send out emails when network issues are discovered.
Fanni Indra Kusuma., [4] The researchers used SNMP to develop a monitoring system for applications based on the syslog messages logged by The Dude. This research is looking to discover whether SNMP monitoring is effective.
Anggar Prabawati., [5] The server monitoring system is built with Hypertext Preprocessor (PHP), Simple Network Management Protocol (SNMP), and SMS Gateway. As a result, administrators can examine the condition of existing servers even when they are not present in the IT room.
Danur Wijayanto., [6]. The information technology management section at Diponegoro University's Faculty of Mathematical Sciences administers network devices such as routers, switches, servers, and access points. Because the administrator manages a large number of devices, the process is carried out manually, relying on client reports if there are network issues.
Alehandrew Michael Fanggidae, Hendi Hermawan, Heny Ispur Pratiwi., [7] The goal of this research is to create server monitoring apps in Python, PHP, and shell script using the SNMP protocol connection. The application is designed to automate the restart of a failed server.
Muhammad Arief Ismaidi., [8] Creating a monitoring system web application that leverages the Simple Network Management Protocol (SNMP) to aid administrators in real time by storing and analyzing server monitoring results.
Vivekanand Mathapati., [9] To acquire accurate results, easily integrate many sensor sources and human system administrator analysis. Human system administrators frequently monitor only a subset of the available system data, such as server load, network load, I/O load, and intrusion detection data. A human analyst may select to further study the behavior of anyone system/server by integrating this information. A resource data analysis system is a technology that enables the quick and easy integration of sensor resources and uses custom-built models to find evidence of processes occurring in the observed environment.
Karyn Vusvyta., [10] Creating a website interface with the assistance of a database, network mapping, and an early warning system (urgent notice) such as Short Message Service.
Susilawati, Fitzgerald dan Renaldy., [11] The goal of developing this application is to create a network monitoring application that can assist network administrators in monitoring the flow of data on the server using a Short Message Services server application that runs on a GSM network, allowing network administrators to monitor the network at any time and from any location.
Nobertus Luan., [12] In order to assist network administrators, a web-based network monitoring application with email notifications will be developed in order to retrieve and save values in the database. This way, information reports that include the availability of network devices are possible.
The difference between this paper than above is this application store queries and resource usages that can be used to help owner make the correct decision. Second, this web application can monitor all web application that already using Java Melody. and easy to add more web application server that want to be monitor, just add the data on JSON.
The research methodology consists of six steps, they are data collection, data analysis, present condition, application design, application development using Kanban method, submission and evaluation. As shown in Figure 1 below, the first step is data collection where data will be got from an interview with the People in Mobile Division at X company. The second step is data analysis where we analyze data that we got after finished step 1 to research and determine what feature that can be implemented and what is not. In step three check workflow from the present condition and from there determine what can we do to lighten developer work at X company for doing server maintenance. In step four, after knowing what feature wants to be implemented and see what can we do to lighten the work of developer, then we made a system planning like Class Diagram, Activity Diagram and Entity Relationship Diagram. Step five, develop a monitoring application using Kanban method. Lastly, in step six submit a monitoring application to get an evaluation from users. If there is a bug, then they would make new Kanban cards so we can fix it right away. And if there is no problem then the application is already completed.
3.1 Data Collection
First we need to interview the production head section, product head section, and product integration head. After got the data that the writer needs, then writer decides to using Kanban method to develop web application monitoring. The reason for using Kanban method for development can be seen below [13].
- Panoramic View
- Defining bottleneck
- Empowered Teams
- The order of features release
- All attention to the most important
- Being focused on work
- Flexibility
- Ability to know everything
- No need to assess features
- Less talk and more work
- Cohesion
- More frequent mistakes
- More flows
- Focus on one task
3.2 Data Analysis
After getting the requirement from step one, then next is to determine what feature that can be implemented and what tools can be used. After discussion with people from company X, we know that all client web app servers already had JavaMelody installed. JavaMelody is an open source application to monitoring JavaEE application. The goal of JavaMelody is to monitor Java or Java EE applications in QA and production environments. It is not a tool to simulate requests from users, it is a tool to measure and calculate statistics on real operation of an application depending on the usage of the application by users [14].
Then after that we explore more about other tools that can be used to display graph in web application. After research for a while, writer found couple of applications that can be used to display graphs on web application and the writer finally choose Grafana. Grafana is an open source based visualization and analysis software. This makes it possible to query, visualize, alert, and explore metrics wherever they are stored [15]. But there is a problem because Grafana cannot access JavaMelody directly, so we need an application that acts as a middleman and that application name is Prometheus.
Prometheus is an open source system for monitoring and alerting created by SoundCloud. Since its inception in 2012, many companies and organizations have adopted Prometheus [16]. Prometheus can read API from web application monitoring that got data from JavaMelody and Grafana can read data from Prometheus to make a graph that can be used to show on web application monitoring. With this research then data analysis is already completed.
3.3 Present Condition
Currently the workflow is like this, Mobile production developer at X company needs to do regular visits or maintenance to check the health of the webapp server and record the daily use of webapp server resources in the form of CPU, Memory Usage, Active Thread, Transaction Per Minute and Garbage Collector every month. Now to record the daily usage of webapp server resources, the developer needs to record it manually by looking at JavaMelody graph and decide what is the average usage resources. This process wastes times a lot between 1,5 to 2 hours per server and because record data is still manual, it will susceptible to mistake on recording data. Also, every developer had different interpretations on reading webapp server resources graph.
3.4 Application Design
From data that already had been collected, then make a class diagram, activity diagram, entity relationship diagram. As shown on Figure 2 below is the web application java application server monitoring class diagram image.
Activity Diagram for web application monitoring can be seen in Figure 3 below
1. Home
When the service is up, application will read JSON file. Inside JSON file have a list of servers that want to be monitored. The application then will hit server and got the status of the server as shown on Figure 4. If green means up, yellow means danger (usage resources server past the limit) or red means server is down.
2. Server Detail
When user clicks server name on the homepage, then the page would be directed to a server detail page. On here, user can see server usage resources from today date at 00:00 until 23:59
3. Search Server Detail by Date
Just like Activity Diagram Server Detail, the difference is the user input date range they want to saw. If already input date and click apply button, application would show usage resources graph according to selected date range as shown on Figure 5 above.
Entity Relationship Diagram (ERD) for web application monitoring can be seen in Figure 6 below
3.5 Application Development Using Kanban Method
As the name implies, the development of web application server monitoring is using Kanban method. List of Kanban cards and their estimation (days) for each card can be seen in table 2.
Table 2. Kanban Card
|
ID |
Kanban Card |
Estimation (Days) |
|
1 |
Research on how to use JavaMelody |
2 |
|
2 |
Research how to make an API that can be read by Prometheus |
2 |
|
3 |
JSON format to use to add or subtract list server |
1.5 |
|
4 |
Set up Java Maven Project and SQL Server 2012 |
1.5 |
|
5 |
Develop homepage |
2 |
|
6 |
Store data usage resource to database |
7 |
|
7 |
Store query that already passed the parameter in the database |
2.5 |
|
8 |
Send an email to the developer if webapp client server usage resources are greater than limit |
2.5 |
|
9 |
Develop Server Detail page to show a graph of usage resources |
7 |
|
Total |
28 |
|
And for Figure 7, 8, 9, 10 and 11 below are Kanban process.
3.6 Submission and Development
Results of black box testing for Server Monitor application can be seen in table 3.
Table 3 Black Box Testing
|
No |
Testing Scenario |
Expected Result |
Test Result |
Conclusion |
|
1 |
Server detail page will open when client name is clicked on homepage |
Display server detail page and show a graph of usage resources |
Successfully displays the server detail page and show a graph of usage resources |
Success |
|
2 |
If there is a connection lost or disconnected |
On Homepage will have a text “There is no internet connection” |
Successfully displaying a text “There is no internet connection” when internet disconnected |
Success |
|
3 |
Search graph usage resources between a date range |
Displays graph usage resources according to the specified date range. |
Successfully displays detail server page according to the specified date range. |
Success |
|
4 |
Click text “back” on server detail page |
Display home page |
Successfully display homepage |
Success |
|
5 |
Click server name on the top of server details page |
Will reset date search range and display today's graph at 00:00 to 23:59 |
Successfully display today's graph at 00:00 to 23:59 |
Success |
|
6 |
If the webapp server is down |
Background color of the name of that server on homepage would change to red and down status on homepage would increase by 1 |
Successfully changes background color to red and the down status increases by 1 on home page |
Success |
|
7 |
If the usage resources server pass the limit |
Background color of the name of that server on homepage would change to yellow and danger status on homepage would increase by 1 |
Successfully changes background color to yellow and the danger status increases by 1 on home page |
Success |
|
8 |
If webapp server monitor time is between maintenance start and maintenance end |
Background color of the name of that server on homepage would change to gray and danger status on homepage would increase by 1 |
Successfully changed the color to gray and the maintenance status increased by 1 on home page |
Success |
|
9 |
If the usage resources server pass the limit several times in a row |
Application will send an email regarding the web app server's usage resources to the developer email address |
Application successfully sent an email alert regarding the web app server's usage resources to the developer email address |
Success |
|
10 |
If query execution time exceeds the max or mean limit |
Application will store that query into database |
Successfully stored query into the database |
Success |
4.1.1 Application Implementation
To run this web application monitoring, server or computer need the minimum and optimum hardware and software as follows:
minimum hardware specification can run web server monitoring that monitors about 35 web applications in 15 to 25 seconds. But if there is a slow response from webapp client, we set the timeout to 20 seconds. Hence, we need about 35 to 45 seconds (include timeout) to run each loop. For the application to run smoothly, we have a threshold to finish each loop below 60 seconds, because the application hits webapp client server every 60 seconds. Therefore, we need a minimum hardware specifications as shown in Table 4.
Table 4 Hardware Specification
|
RAM |
4 GB |
|
Memory |
100 MB |
|
CPU |
4 x 2.20 |
Minimum and optimum software specification is show on Table 5. The reason Prometheus and Grafana need a correct version is on the Grafana documentation, it states that the scripted dashboard that needed to make a graph so web application server monitoring can show that graph is deprecated and will be removed in the next release of Grafana. For that specific reason, we use Grafana v6.0.2 which the scripted dashboard has not yet deprecated, and then we use Prometheus v2.8.0 because it is released in about the same period as Grafana v6.0.2 so minimizing if there is an incompatible version.
Table 5 Software Specification
|
Database |
SQL SERVER 2012 |
|
Javamelody |
> v 1.77.0 |
|
Prometheus |
v 2.8.0 |
|
Grafana |
v 6.0.2 |
1.1.2 Application Guidelines
Here is the guideline to run the app. On the home page, the section at the top contains text like current Server Monitor and Server Status, which is Maintenance, Up and Down, with total count of the web server that is related for each of that status. The explanation of each status is as follows:
- Maintenance Status which has a gray color indicates that the webapp server is under maintenance.
- Up Status which has a green color indicates that the webapp server is up and can be accessed by the users.
- Down Status which has a red color indicates that the webapp server is down and cannot be accessed by the users. However, this could also happen if the web server monitoring did not get a response from the web app server in about 20 seconds (timeout).
- Danger Status which has a yellow color indicates that the resource consumption of the web app server is above the limit that has been set.
Figure 12 below is the list of the web app server with their status. If we click one of the servers, system will redirect us to the server detail page.
If there is no connection, system will not show anything but a text “There is no internet connection “will be shown as mentioned on Figure 13.
Server detail page will be open when the user clicks one of the servers on the home page. Here, it has the name of the webapp in the top middle section. In the top right section, it has a back button that will redirect us back to the home. Below as shown on Figure 14, the name section, it has an input section to time filter period.
In addition, this page also contains 6 graphs that represent resource usage of the web app server. Below is the explanation of the 6 graph.
1. Active Threads
Active Threads represent processes that are running in the webapp server.
2. System CPU
System CPU represents CPU usage of the server in percentage metrics. The higher the usage indicates that we need to either improve the code and query to minimize the CPU usage, or upgrade our hardware.
3. Physical Memory
Physical Memory represent memory usage by OS (Operating System)
4. Used Memory
Used Memory represent RAM (Random Access Memory) usage in the server.
5. Garbage Collector
Garbage Collector is a process to remove unused objects. In Java programming language, Garbage Collector runs automatically until the app is closed. This Graph is important to know if there is a memory leak that can lead to an error out of memory.
6. Free Disk Space
Free Disk Space is the remaining space in the server disk. The lower the graph gets, indicates that we need to remove files that are not used like log or move the file to another disk.
We can show graph data in our desired period by input date, hour and minutes in From and End section and then click the button apply. Here is an example of the graph in the range of 30 minutes as shown on Figure 15.
Metrics Page (API)
This page providing data about metrics so Prometheus can read this API and make data that is readable by Grafana.
System CPU and garbage collector unit measure in percentage, while used memory, physical memory and free disk space unit measure in a byte. For active thread, it will measure with its actual usage thread that in the process when the data is received. After Prometheus reads the API, then Grafana will use the API from Prometheus to create a graph that can be displayed in web app server monitoring. Below is an example of API that Prometheus read.
# HELP CSFMOSS_web_P_last_value_system_cpu_load CSFMOSS_web_P value per minute
# TYPE CSFMOSS_web_P_last_value_system_cpu_load gauge
CSFMOSS_web_P_last_value_system_cpu_load 0.0
# HELP CSFMOSS_web_P_last_value_used_memory CSFMOSS_web_P value per minute
# TYPE CSFMOSS_web_P_last_value_used_memory gauge
CSFMOSS_web_P_last_value_used_memory 0.0
# HELP CSFMOSS_web_P_last_value_gc CSFMOSS_web_P value per minute
# TYPE CSFMOSS_web_P_last_value_gc gauge
CSFMOSS_web_P_last_value_gc 0.0
# HELP CSFMOSS_web_P_last_value_active_threads CSFMOSS_web_P value per minute
# TYPE CSFMOSS_web_P_last_value_active_threads gauge
CSFMOSS_web_P_last_value_active_threads 0.0
# HELP CSFMOSS_web_P_last_value_used_physical_memory_size CSFMOSS_web_P value per minute
# TYPE CSFMOSS_web_P_last_value_used_physical_memory_size gauge
CSFMOSS_web_P_last_value_used_physical_memory_size 0.0
# HELP CSFMOSS_web_P_last_value_free_disk_space CSFMOSS_web_P value per minute
# TYPE CSFMOSS_web_P_last_value_free_disk_space gauge
CSFMOSS_web_P_last_value_free_disk_space 0.0
# HELP CSFMOSS_ws_P_last_value_system_cpu_load CSFMOSS_ws_P value per minute
# TYPE CSFMOSS_ws_P_last_value_system_cpu_load gauge
CSFMOSS_ws_P_last_value_system_cpu_load 17.250841529191895
# HELP CSFMOSS_ws_P_last_value_used_memory CSFMOSS_ws_P value per minute
# TYPE CSFMOSS_ws_P_last_value_used_memory gauge
CSFMOSS_ws_P_last_value_used_memory 1.477268976E9
# HELP CSFMOSS_ws_P_last_value_gc CSFMOSS_ws_P value per minute
# TYPE CSFMOSS_ws_P_last_value_gc gauge
CSFMOSS_ws_P_last_value_gc 0.0
# HELP CSFMOSS_ws_P_last_value_active_threads CSFMOSS_ws_P value per minute
# TYPE CSFMOSS_ws_P_last_value_active_threads gauge
CSFMOSS_ws_P_last_value_active_threads 2.0
# HELP CSFMOSS_ws_P_last_value_used_physical_memory_size CSFMOSS_ws_P value per minute
# TYPE CSFMOSS_ws_P_last_value_used_physical_memory_size gauge
CSFMOSS_ws_P_last_value_used_physical_memory_size 2.4339001344E10
# HELP CSFMOSS_ws_P_last_value_free_disk_space CSFMOSS_ws_P value per minute
# TYPE CSFMOSS_ws_P_last_value_free_disk_space gauge
CSFMOSS_ws_P_last_value_free_disk_space 7.6714299392E10
List Server Addition
If the user wants to add or remove web app server they can change JSON data that can be read by the app. This method is preferable because it is easier to understand and modify.
To add or remove the server on the home page user can add or remove object data in the JSON file with format as follows.
{
"Server": {
"name" : "xxx", (name that will be displayed on the web app server monitor)
"url" : "https://www.xxx.com/ monitoring", (app Url that will be monitored and has javamelody applied)
"servertype" : "webapp", (Can be choose between webapp / services)
"username": "username", (username to access javamelody from the given url)
"password": "password", (password to access javamelody from the given url)
"sqlmean": "6500", (average time (in miliseconds), which if there is a sql query that runs above the limit of the sqlmean, then the system will record it in the database)
"sqlmax": "7500", (max time (in miliseconds), which if there is a sql query that runs above the sqlmax, then the system will record it in the database)
"alertThreads": "10", (Maximum thread in which if there is thread usage that exceeds the limit of activeThreads, then the web app status will change color to yellow or danger)
"alertFreedisk": "150000000000", (If the current free disk (in bytes) is below the AlertFreeDiskValue, then the web app status will change color to yellow or danger)
"alertGcmax": "5", (Maximum GC (in percentage). if GC usage volume exceeds alertGcmax, then the web app status will change color to yellow or danger)
"alertGcmean": "3", (average GC (in percentage), if GC average usage exceeds alertGcmean value, then the web app status will change color to yellow or danger)
"alertSystemCPU": "50", (Maximum SystemCPU (in percentage), if SystemCPU usage exceeds alertSystemCPU value, then the web app status will change color to yellow or danger)
"alertPhysicalMemory": "16000000000", (Maximum PhysicalMemory (in bytes), if PhysicalMemory usage exceeds alertPhysicalMemory value, then the web app status will change color to yellow or danger)
"alerttime": "5", (if there is a web app server that exceeds n number times alert in a row, then the system will send notification / email to the production team to check the server condition)
"alertDowntime": "120", (If email had been sent to the receiver and even if it exceeds n number times alert, then the email will not be sent, except if the period from the last email sent to current time has exceeded AlertDownTime (in minutes). This will minimize spam email to the team if the server is continuously down and exceed the alert limit)
"maintenancestart": "00.00", (if the current time is in the range of maintenancestart and maintenanceend, then the web app status will change color to grey or maintenance)
"maintenanceend": "00.00"
}
}
Email Alert
This feature is used to notify the production team developer if there is a server that down or resource usage is past the limit that had already been set on JSON data that can be seen above.
1. Server Down Alert
An Email will be sent to the developer when the web app server cannot be accessed by server monitoring or web application got a timeout when hits webapp client as shown on Figure 16.
2. Query Alert
An Email will be sent to the developer when there are queries that past mean or max time that already had been set on JSON as shown on Figure 17.
3. Alert Free Disk Space
An Email as shown on Figure 18 will be sent to the developer when the free disk space is below the parameter that had been set on JSON.
Information
Alert GC Mean, GC Max, Physical Memory, System CPU, Threads have the same format as Alert Free Disk Space. The difference is free disk below the parameter then application sent an email to developer, but for anything else beside the free disk, if usage resources above parameter then application sent email to the developer.
Database
In the database there are a total of 6 tables. Name and description of each table can be read as follows.
This is used to send emails to developers if there is a webapp server that experiences down or resource usage that exceeds the limit that had been set.
This table has 4 columns. which is:
1. Emailid
Emailid is the primary key.
2. Email_to
Email_to is what email address of the developer that will receive the email alert. If there is more than 1 email address that wants to be sent. Then use semicolon (;) as a separator. Example if want to send email to 2 email address: [email protected];[email protected]
3. Email_cc
Email_cc is the email address that will get alert email for carbon copy. If there is more than 1 email address that wants to get alert email for carbon copy. Then can use semicolon (;) as a separator. Example to send email to two recipients to get the carbon copy: [email protected];[email protected]
4. Is_active
To mark if the email_to or email_cc is still active. If not active, email will not be sent to the recepient.
- Path Data
PathData Table has 2 columns, which is:
1. Directory
Path in which JSON or Image file will be read by the application.
2. FileType
Inside this only has 2 data, which are json and img.
- Query Data
This table contains 8 columns which are stored a Query that running past the max or mean limit, each column can be read below:
- QueryID (PK)
- ClientCode
- URL
- QueryString: Heavy Query that is running on client webapp.
- MeanTime: Average time needed to execute QueryString above.
- MaxTime: Maximum time needed to execute QueryString above.
- FlagQuery: Value only have 1 or 0 where 1 is mean that query using operator LIKE
- TimeStamp: Time when QueryString is stored on the database.
- RV_ReportV2
This table contains 7 columns which is used to store webapp client usage resources every minute, each column can be read below:
- ID (PK)
- ClientCode
- Max_Memory: Memory usage on webapp client when data is taken.
- Max_Thread: Thread usage on webapp client when data is taken.
- Max_TPM: Sum transaction per minute on webapp client when data is taken.
- Time_Stamp: Datetime when data is taken.
- Max_GC: Garbage Collector usage on webapp client when data is taken.
- RV_ReportSummary
In RV_ReportSummary Table, it contains daily summary of server usage resources from every webapp client server. Data from here is created by stored procedure at every start of the month. There are 8 columns of this table which is time data collected, name of the server, avg memory, max memory, max thread, avg TPM, max TPM, max GC on each day for each client.
- Utilities
In Utilities Table there are 4 columns as follows:
1. ID
Primary key.
2. Name
There are 7 data in this column which is:
- Sql_start_check: Start hour for application begin to check sql query and send alert if there is a query that passes mean or max limit.
- Sql_pause_check: interval in an hour between each check.
- Sql_end_check: End hour for application to end check sql query.
- Timeout: ping duration to the webapp server (ms).
- Send_sql_email: 1 or 0 (1 means application will send email to the developer if there is query that exceeds the limit or using operator LIKE)
- Send_email_down: 1 or 0 (1 means application will send email to developer if there is webapp client server which cannot be reached by application or got timeout response many times in a row)
- Alert_count_threshold: How much in a row webapp server got status down before sending email to the developer.
3. Value
Value of the name above. To see the example of the value, one can refer to the name section above.
4. Description
Description of the name above. To see the example of the value, one can refer to the name section above.
Table comparison evaluation between JavaMelody with web app server monitoring.
Table 6 provided the comparison data between JavaMelody that is taken manually by seeing and interpreting the value of the graph each day with web server monitoring for usage memory in Megabytes and percentage. For percentage we can use the formula as follows:
- If Manual <= Web App Monitoring, then:
- If Web App Monitoring < Manual, then:
Table 6 Comparison data between JavaMelody
|
Server 1 Webapp |
|||||
|
18 June 2021 |
AVG Memory |
MAX Memory |
Max Thread |
AVG TPM |
MAX TPM |
|
Manual |
2000 |
3300 |
0 |
66 |
331 |
|
Server Monitoring |
1795.585 |
3287.001 |
0 |
65 |
394 |
|
Accuracy |
90% |
100% |
100% |
98% |
84% |
|
19 June 2021 |
|||||
|
Manual |
1900 |
3200 |
0 |
32 |
245 |
|
Server Monitoring |
1750.147 |
3244.793 |
0 |
31 |
284 |
|
Accuracy |
92% |
99% |
100% |
97% |
86% |
|
20 June 2021 |
|||||
|
Manual |
1800 |
3250 |
0 |
39 |
225 |
|
Server Monitoring |
1743.984 |
3256.274 |
0 |
38 |
277 |
|
Accuracy |
97% |
100% |
100% |
97% |
81% |
|
Server 1 Webservices |
|||||
|
18 June 2021 |
AVG Memory |
MAX Memory |
Max Thread |
AVG TPM |
MAX TPM |
|
Manual |
4000 |
9300 |
10 |
1200 |
4200 |
|
Server Monitoring |
3816.403 |
9639.635 |
10 |
1213 |
4587 |
|
Accuracy |
95% |
96% |
100% |
99% |
92% |
|
19 June 2021 |
|||||
|
Manual |
4000 |
10000 |
8 |
868 |
3848 |
|
Server Monitoring |
3622.263 |
10166.94 |
8 |
868 |
4544 |
|
Accuracy |
91% |
98% |
100% |
100% |
85% |
|
20 June 2021 |
|||||
|
Manual |
2000 |
3500 |
2 |
41 |
199 |
|
Server Monitoring |
1911.265 |
3629.714 |
2 |
40 |
202 |
|
Accuracy |
96% |
96% |
100% |
98% |
99% |
Web application server monitoring has been made to help the work of developer that wants to maintenance multiple Java EE server Applications using one application and has been done as well as possible. Web applications are built by prioritizing data accuracy and ease of use of web applications. Then the web application server monitoring can be concluded as follows:
-
Web application server monitoring can store webapp client usage resources.
-
Web application server monitoring can show graph usage resources in each server.
-
Web application server monitoring can send email to recipient when there is usage resource pass the parameter.
This monitoring web server application also still has several shortcomings, namely:
-
If the web server client is down or web server application monitoring got timeout when hit webapp client. It will make accuracy worse.
-
UI that is not attractive enough because only use graph and color.
-
There is no login feature, so the list server can be added or subtract by anyone who can access the office's internal server.
Acknowledgements : We would like to acknowledge the support and facilities given by Binus Graduate Progam as well as Advanced Institute of Industry Technology, Tokyo, Japan during the research.
Author contributions : SS and FLG has initiated the study, conducted the experiments, and made the first draft of this paper. TM gave experts’ opinions and contributed to the experimental phase. All authors have read and approved the final manuscript.
Funding : Not applicable.
Availability of data and materials: The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.
Ethics approval and consent to participate: Not applicable.
Consent for publication : Not applicable.
Competing interests : The authors declare that they have no competing interests.
Author details : 1Computer Science Department, Binus Graduate Program, Binus University, Jakarta 11480, Indonesia;. 2Computer Science Department, BINUS Graduate Program – Doctor of Computer Science, Binus University, Jakarta 11480, Indonesia, 3Advanced Institute of Industry Technology, Tokyo, Japan.
- Argaez Enrique De. WORLD INTERNET USAGE AND POPULATION STATISTICS. 2021 Year-Q1 Estimates. Available online: https://www.internetworldstats.com/stats.htm (accessed on 24 September 2021).
- Asosiasi Penyelengara Jasa Internet Indonesia. Laporan Survei Internet APJII 2019 – 2020 [Q2]. 2020. Available online: https://apjii.or.id/survei (accessed on 22 July 2021).
- Ray, Endang. Pengembangan Aplikasi Monitoring Server Berbasis Mobile Web Dengan Sistem Notifikasi Email. (2015) Available online: http://repository.uinjkt.ac.id/dspace/bitstream/123456789/28103/1/ENDANG%20RAY-FST.pdf (accessed on 07 March 2021).
- Kusuma, Fanni Indra. Perancangan Sistem Monitoring Perangkat Jaringan Berbasis SNMP. 2015. Available online: http://eprints.ums.ac.id/38600/27/Naskah-Publikasiku.pdf (accessed on 07 March 2021).
- Prabawati, Anggar. Analisis Dan Perancangan Pemantau Jaringan Server Menggunakan PHP, SNMP Dan SMS Gateway Pada PT. PLN (Persero) Rayon Ponogoro. 2013. Available online: http://eprints.umpo.ac.id/495/2/KusumaHAL%20DEPAN%2C%20BAB%20I%2C%20DAFPUS%20anggar.pdf (accessed on 07 March 2021).
- Wijayanto Danur. Aplikasi Monitoring Perangkat Dan Aktivitas Pengguna Pada Jaringan Menggunakan Protocol SNMP Dan Squid Proxy. 2016. Available online: https://core.ac.uk/download/151235843.pdf (accessed on 19 March 2021).
- Fanggidae, A. M., Hermawan, H., & Pratiwi, H. I. Sistem Monitoring Server Dengan Menggunakan SNMP. 2019. Available online: http://www.ojs.upj.ac.id/index.php/journal_widya/article/view/218 (accessed on 20 July 2021).
- ISMAIDI, M. A. Pengembangan Aplikasi Monitoring Performa Server Menggunakan Simple Network Management Protocol. 2013. Available online: http://etd.repository.ugm.ac.id/penelitian/detail/65035 (accessed on 19 July 2021).
- MATHAPATI, V. Performance Analysis Of System Resources By Server Monitoring 2013. Available online: http://www.ijirset.com/upload/july/30A_PERFORMANCE.pdf (accessed on 19 July 2021).
- VUSVYTA, K. Design And Implementation Fast Response System Monitoring Server Using Simple Network Management Protocol (SNMP). 2017. Available online: http://edocs.ilkom.unsri.ac.id/1654/1/TASK2_Manajemen%20Jaringan_SNMP.pdf (accessed on 07 March 2021).
- Susilawati, S., Fitzgerald, J., & Renaldy, C. Analisis dan Perancangan Aplikasi Traffic Monitoring Server Menggunakan SMS Pada PT. Anugrah Catur Abadi. 2008. Available online: http://eprints.binus.ac.id/2665/1/2008-2-00210-IF%20Abstrak.pdf (accessed on 19 July 2021).
- Luan, N. Aplikasi Monitoring Jaringan Berbasis Web Dengan Notifikasi Email. 2019. Available online: http://eprints.mercubuana-yogya.ac.id/5356/ (accessed on 19 July 2021).
- Ku, Pavel. 14 Reasons To Choose Kanban For Software Development. 2018. Available online: https://medium.com/hygger-io/14-reasons-to-choose-kanban-for-software-development-d3ddb420d273 (accessed on 28 March 2021).
- JavaMelody. JavaMelody : monitoring of JavaEE applications. 2019. Available online: https://github.com/javamelody/javamelody/wiki (accessed on 09 August 2020).
- Grafana. What is Grafana. Available online: https://grafana.com/docs/grafana/latest/getting-started/what-is-grafana/ (accessed on 09 August 2020).
- Prometheus. What is Prometheus?. Available online: https://prometheus.io/docs/introduction/overview/ (accessed on 24 September 2021)