• another investigation was being done, why the server suddenly behaved strange / dead. Apparently, a 114GB log file was growing, that caused server wide issues. Well, apparently the EPM events run in a file descriptor leak. One line of code fixed that.
• Today, I saw that the CPU was running non stop at around 50%, which should not happen either. Serious issues were found in that regard (db, EPM events) and got fixed too. Overall the result:

TL;DR: PostgreSQL "idle" processes: 66-64% → <1% CPU PHP epm_events scripts: 14% → 2-5% CPU each docker-proxy: 40% → <5% CPU (should auto-improve)

Total CPU reduction: ~170% → ~10-15%

This should make HWS Connect more responsive for you again.

• another investigation was being done, why the server suddenly behaved strange / dead. Apparently, a 114GB log file was growing, that caused server wide issues. Well, apparently the EPM events run in a file descriptor leak. One line of code fixed that.
• Today, I saw that the CPU was running non stop at around 50%, which should not happen either. Serious issues were found in that regard (db, EPM events) and got fixed too. Overall the result:

TL;DR: PostgreSQL "idle" processes: 66-64% → <1% CPU PHP epm_events scripts: 14% → 2-5% CPU each docker-proxy: 40% → <5% CPU (should auto-improve)

Total CPU reduction: ~170% → ~10-15%

This should make HWS Connect more responsive for you again.