AppMon Message Protocol Specification

1. Server-to-Client Message Structure (Push/Broadcast)

All data messages sent from the server to the client consist of four primary fields using a colon (:) as a delimiter.

{instanceName}:{exporterType}[/{subType}]:{exporterName}:{content}

1.1 Field Definitions

1.2 Client-side Parsing and Mapping Rules

1. Field Separation: Split the message based on the colon (:) delimiter. Since the content may contain colons, locate only the first three indices first.
2. Type Analysis: Check the exporterType field for the / character to separate the base type and the subType.
3. Identifier Generation (exporterKey): Combine the fields into instanceName:pureType:exporterName to create the exporterKey. This key serves as a unique identifier to map data to UI elements (console boxes, charts, etc.).
4. Data Processing: Perform appropriate UI logic (e.g., prepend, append, or update chart) based on the subType.

2. Specification by Exporter Type

2.1 Log (Log)

Handles real-time log streaming and historical log data. The content is plain text.

• Real-time Log (LOG): No specific sub-type.
  • Format: server1:log:root.log:Log message content...
• Previous Log (PREVIOUS): Uses the p sub-type.
  • Format: server1:log/p:root.log:Log message content...
  • The client recognizes this to prepend the data to the top of the console box.

2.2 Visualization Data (Data)

Transmits statistical data for rendering charts and graphs.

• Chart Data (CHART): Uses the chart sub-type; content is in JSON format.
  • Format: server1:data/chart:activity:{"labels":[...], "data1":[...], "rolledUp":false}
  • Provides a flattened JSON structure for direct data access.

2.3 Event (Event)

Transmits system status changes or user session events. Content is in JSON format.

• Format: server1:event:session:{"numberOfActives":10, "createdSessions":[...], ...}

2.4 Metric (Metric)

Transmits single-value performance indicators. Content is in JSON format.

• Format: server1:metric:jvm.memory:{"format":"{used}/{max}", "data":{"used":"512MB", "max":"2GB"}}

3. Client-to-Server Command Structure (Pull/Request)

Used when the client requests a specific action from the server. Based on the APON (Aspectran Object Notation) format, consisting of a set of key:value pairs delimited by semicolons (;).

The CommandOptions class on the server parses this string into an object and determines the action to perform based on the value of the command key.

3.1 Command Format

command:{commandName};{optionKey1}:{value1};{optionKey2}:{value2}

3.2 Key Commands and Option Parameters

3.3 Command Examples

• Requesting Previous Logs: command:loadPrevious;instance:appmon;logName:root.log;loadedLines:500 (Description: Request data prior to the 500 lines already loaded for root.log on the appmon instance.)

• Refreshing Specific Instance Data: command:refresh;instance:server1;dateUnit:hour (Description: Request re-aggregated data for the server1 instance on an hourly basis.)

• Joining a Monitoring Session: command:join;instancesToJoin:server1,server2;timeZone:Asia/Seoul (Description: Join monitoring groups for two instances and provide client time zone info.)

4. Extension Guide

To add a new data format or state:

1. Select the most appropriate base type from the existing exporterType list.
2. Define a new subType after the / character if necessary.
   • e.g., log/e for categorizing error logs.
   • e.g., event/alert for real-time alert notifications.
3. Add the corresponding processing logic for the new subType in the client’s processMessage dispatch logic.