APM infrastructure agent - GCP extension regularly disconnects

Products

• DX Application Performance Management
• DX APM SaaS
• CA Application Performance Management Agent (APM / Wily / Introscope)

Issue/Introduction

APM infra agent running with GCP extension and connected to on-prem DX OI. 

APM infra agent keeps shutting itself down. 

APM infra agent with GCP extension shuts-down itself...

10/28/21 06:08:38 PM CEST [INFO] [IntroscopeAgent.AccControllerManagementService] Executing ACC Controller Process shutdown hook
10/28/21 06:08:39 PM CEST [WARN] [IntroscopeAgent.IsengardServerConnectionManager] Introscope Enterprise Manager apmservices-myserver:80,com.wily.isengard.client.transport.websocket.WebSocketFactory requested disconnect
10/28/21 06:08:40 PM CEST [WARN] [IntroscopeAgent.IsengardServerConnectionManager] Lost contact with the Introscope Enterprise Manager at apmservices-myserver:80.
10/28/21 06:08:40 PM CEST [INFO] [IntroscopeAgent.AccControllerManagementService] Successfully stopped ACC Controller Process
10/28/21 06:08:40 PM CEST [INFO] [IntroscopeAgent.InfrastructureAgent] AutoTracingTriggerServiceBean passivated.
10/28/21 06:08:44 PM CEST [INFO] [IntroscopeAgent.IsengardMessaging] Isengard Input Stream closed.

Why does this occur?..10/28/21 06:08:39 PM CEST [WARN] [IntroscopeAgent.IsengardServerConnectionManager] Introscope Enterprise Manager apmservices......  .WebSocketFactory requested disconnect   => why the server requests a disconnect?

 

Environment

Release : 21.3

Component : APM Infrastructure gent

Cause

wrapper log showed agent was failing with out of memory errors -

INFO   | jvm 5    | SLF4J: See http://www.slf4j.org/codes.html#StaticLoggerBinder for further details.
INFO   | jvm 5    | java.lang.OutOfMemoryError: GC overhead limit exceeded
STATUS | wrapper  | The JVM has run out of memory.  Restarting JVM.

 

Resolution

Increased heap to 2Gb

Recommended minimum heap size for APMIA with GCP, Azure or AWS extensions is 2Gb