We use servicecontrol to monitor more or less 15 endpoints. In production(ravendb compacting has been done), the process uses around 4 GB.
We don’t have a high workload: 500000 messages in the past 30 days. The retention period is 30 days and 15 days for archived messages.
We have only one custom check, which is the dead letter queue monitor based on your sample.
The heartbeat is configured by default as 10 seconds.
Is it the expected memory usage?
Moreover, I attached the debugger to the process and it’s constantly throwing exceptions:
Is it ok?