Should we restore thread pools? It was established that they are not responsible for the recent outages. Note that interrupting timed out threads should mitigate the issue thread pools are supposed to solve.
If we do restore thread pools, then we should ensure clean pool shutdown when application is undeployed (#350 doesn't do this, because it doesn't know application lifecycle). There are edge cases where this can cause problems and is especially dangerous during outage, when admin acts fast, not necessarily following all procedures, and the java.lang.OutOfMemoryError can send his investigation in a wrong direction. It happened to me in stage.
Should we restore thread pools? It was established that they are not responsible for the recent outages. Note that interrupting timed out threads should mitigate the issue thread pools are supposed to solve.
If we do restore thread pools, then we should ensure clean pool shutdown when application is undeployed (#350 doesn't do this, because it doesn't know application lifecycle). There are edge cases where this can cause problems and is especially dangerous during outage, when admin acts fast, not necessarily following all procedures, and the java.lang.OutOfMemoryError can send his investigation in a wrong direction. It happened to me in stage.