AC 5.x Monitoring Parameters
- Rajeev Rai
This document covers the list of parameters that we should monitor to make sure that Adeptia Connect application is running properly in the Kubernetes environment.
In the table below the threshold value is mentioned based on the ideal situation. The threshold value can be changed as per the need.
Monitoring Level | Metrics to monitor | Threshold | Comment |
|---|---|---|---|
Kubernetes | Kubernetes Cluster Availability | Up/Down |
|
No. of total worker nodes | 80% |
| |
Worker nodes CPU Utilization | 90% |
| |
Worker nodes Memory Utilization | 80% |
| |
Worker node disk usage | 80% |
| |
PVC disk usage | 80% |
| |
Pods | CPU usage | 90% of the assigned Limit |
|
Memory Usage | 80% of the assigned Limit |
| |
No. of Evicted Pods | 10 | If pods are evicting frequently, it indicates some misconfiguration. Large number of evicted pods may lead to inaccessible cluster. | |
Application
| Pods of each microservices should be in running state | Number of minimum Replica set for each microservices should be running |
|
Number of pods running for each Microservice | As per the number defined in the Replica Set and the HPA configuration. | If it is increasing beyond a limit, we should investigate, whether the requests are increasing, or we need to configure the resource properly. | |
JVM usage | 80% | The heap usage should be monitored for each Microservices. | |
Number of queued messages in the Rabbit MQ | As per the requirement | If the number of Queued message are increasing beyond the expected level, then we should check whether the Process flow are executing or not. We should also check whether the PFs are taking more time than expected. | |
License expiry | 15 Days |
| |
Database
| Database Connection Count Below environment variables in each Pod defines the number of max active connection to the database: For backend database :
For log database
| Number of active database connection should remain below 80% of some of count defined in the environment variables of Backend and Log databases. | This connection count has to be monitored for each Microservice. In case the number of connections are increasing, we need to check whether we are increasing the load in the application. If the number of connection are increasing due to increased load then we need to adjust the active connection counts in these variables. |
| Diskspace | 80% |
|
| CPU Usage Alert | 80% Usage Alert |
|
| Memory Usage | 80% Usage Alert |
|
|
|
|
|