Not sure. If Lightrun can put data into CloudWatch, then SenseDeep would ingest that and trigger alarms.
Lightrun seems to be pretty intrusive requiring an agent and runtime code modification. The approach we use is 100% external to your Lambda code yes is still dynamic.
Yes it requires an agent but not runtime code modification. The plus side is snapshots, metrics and deep logging capabilities. Sort of like debugging a local app directly in production.
How does it take snapshots without VM or code modification? I've seen this technique where you use the VM debugging API but it is not without performance cost. We use EMF for metrics which works great.
It very much depends on the platform you're in. E.g. in Java it uses the low level native JVMTI API which we invoke from C++. The overhead when this is unused is negligible. You do pay an overhead when it's on but only for the actual snapshot/applicable code. Once the snapshot is grabbed the overhead is gone.
We have code that detects overuse of CPU and throttles down logs/conditions etc.
For further actions, you may consider blocking this person and/or reporting abuse
We're a place where coders share, stay up-to-date and grow their careers.
Not sure. If Lightrun can put data into CloudWatch, then SenseDeep would ingest that and trigger alarms.
Lightrun seems to be pretty intrusive requiring an agent and runtime code modification. The approach we use is 100% external to your Lambda code yes is still dynamic.
Yes it requires an agent but not runtime code modification. The plus side is snapshots, metrics and deep logging capabilities. Sort of like debugging a local app directly in production.
How does it take snapshots without VM or code modification? I've seen this technique where you use the VM debugging API but it is not without performance cost. We use EMF for metrics which works great.
Can you emit your log data into CloudWatch so the SenseDeep viewer & alarms could process? That would be cool.
It very much depends on the platform you're in. E.g. in Java it uses the low level native JVMTI API which we invoke from C++. The overhead when this is unused is negligible. You do pay an overhead when it's on but only for the actual snapshot/applicable code. Once the snapshot is grabbed the overhead is gone.
We have code that detects overuse of CPU and throttles down logs/conditions etc.