Dynatrace Plugin Apigee increase CPU to 100%

[mateoporcar]

Hello,

We installed the following Dynatrace Apigee plugin in our dev environment: https://www.dynatrace.com/hub/detail/apigee/

This plugin exports the apigee analytics to dynatrace dashboards. The issue is that when we activated it the cpu of the linux server where is running the edge-management-server (port 8080) the cpu increased up to 100% and never go down.

We tried filter only by 1 api-proxy, reduce time intervals between collections, reduce the dymensions.. but always the cpu goes to 100%.

The edge-management-server do not show any error, only the following logs:

2024-12-13 11:36:43,496 org:organization-name env:staging target:/v1/organizations/organization-name/environments/staging/stats/api_product%2Crequest_path%2Cclient_id%2Cresponse_status_code%2Capiproxy contextId: action: qtp874111547-20690 INFO  i.a.a.q.u.f.Expression - Expression$Interpreter.asSQLExpression() : Interpreting apiproxy in 'accounts-api','contracts-api','customers-api','inbox-gateway_api','master-data-api','oauthv2','profiler-api'
2024-12-13 11:36:43,496 org:organization-name env:staging target:/v1/organizations/organization-name/environments/staging/stats/api_product%2Crequest_path%2Cclient_id%2Cresponse_status_code%2Capiproxy contextId: action: qtp874111547-20690 INFO  c.a.a.q.i.UAPQueryHelper - UAPQueryHelper.isUAPOnboarded() : Checking if organization-name.staging is present in uap onboarded cache
2024-12-13 11:36:43,975 org:organization-name env:staging target:/v1/organizations/organization-name/environments/staging/stats/api_product%2Crequest_path%2Cclient_id%2Cresponse_status_code%2Capiproxy contextId: action: qtp874111547-20690 INFO  c.a.a.q.i.UAPQueryHelper - UAPQueryHelper.isUAPOnboarded() : Checking if organization-name.staging is present in uap onboarded cache
2024-12-13 11:36:44,205 org:organization-name env:staging target:/v1/organizations/organization-name/environments/staging/stats/api_product%2Crequest_path%2Cclient_id%2Cresponse_status_code%2Capiproxy contextId: action: qtp874111547-20690 INFO  c.a.a.d.impl.BaseResource - BaseResource.buildModel() : path for current request - organizations/organization-name/environments/staging/stats/api_product,request_path,client_id,response_status_code,apiproxy (org:env) is organization-name:staging and query params:{topk=[1000], sortby=[avg(total_response_time)], sort=[DESC], select=[tps,sum(cache_hit),sum(message_count),sum(is_error),sum(target_error),avg(total_response_time),avg(request_size),avg(response_size),avg(target_response_time),avg(request_processing_latency),avg(response_processing_latency)], timeUnit=[minute], timeRange=[12/13/2024 14:26~12/13/2024 14:27], filter=[(apiproxy in 'accounts-api','contracts-api','customers-api','inbox-gateway_api','master-data-api','oauthv2','profiler-api')]}
2024-12-13 11:36:44,207 org:organization-name env:staging target:/v1/organizations/organization-name/environments/staging/stats/api_product%2Crequest_path%2Cclient_id%2Cresponse_status_code%2Capiproxy contextId: action: qtp874111547-20690 INFO  c.a.a.q.i.UAPQueryHelper - UAPQueryHelper.isUAPOnboarded() : Checking if organization-name.staging is present in uap onboarded cache

There is any know issue related to this?

Thanks,

Mateo.

[AlexET]

Hey @mateoporcar, thanks for your question - We’re keeping an eye on this thread to ensure you get the answers you need, and we also invite the community to pitch in with their thoughts.

Your patience means a lot to us. Additionally, consider checking our articles and upcoming events! 😊

[gcpsean]

Hi Mateo,

Which process is consuming the CPU? Is it the Dynatrace process or an Apigee process? Also, can you confirm this is for Edge Private Cloud (OPDK)?

[mateoporcar]

Hello @gcpsean , thanks for the response.

The process consuming the CPU is the Apigee process, the Management Server. 

Yes it is Edge Private Cloud (OPDK). In this image you can see that process consume 97.3%. 

The drop in consumption observed at the end of this table is after turning off the dynatrace plugin.

Thanks,

Mate.

[gcpsean]

Hi Mate,

Unfortunately, we did not write and do not support this plugin. If I had to guess, Dynatrace's plugin is either being too aggressive in polling the Apigee management APIs, causing resource exhaustion; it is doing something which puts the management API process itself into a bad state, such as causing lockfile contention with /opt/apigee/var/run; or the OS is interfering via something like selinux configurations (triggered by the Dynatrace agent) which is blocking access to file resources the Management server requires and sending it into an infinite loop.

I would suggest you raise a support case with Dynatrace to investigate on their side.

Thank you!

[mateoporcar]

Hello @gcpsean  and @AlexET , any update about this?

Thanks,

Mate.