@Wunder - thanks for the info! We are in the process of adding min/max and percentiles to SPM for Solr/ES, etc. Do also have some percentile metrics for top N queries.
Re NewRelic and cores - amazing. In SPM for Solr you can filter by cores, request handler, time, and maybe a few more things.
We use mostly use a home-grown Python script, because we need percentiles (90, 99, 99.9) and number of queries over a certain time (5s, 10s, etc.).
New Relic is good for a quick look, but it does not report separate times for separate cores, and it reports average response time, which is the wrong metric for a skewed distribution (like response times).
@Wunder - thanks for the info! We are in the process of adding min/max and percentiles to SPM for Solr/ES, etc. Do also have some percentile metrics for top N queries.
Re NewRelic and cores - amazing. In SPM for Solr you can filter by cores, request handler, time, and maybe a few more things.
We use mostly use a home-grown Python script, because we need percentiles (90, 99, 99.9) and number of queries over a certain time (5s, 10s, etc.).
New Relic is good for a quick look, but it does not report separate times for separate cores, and it reports average response time, which is the wrong metric for a skewed distribution (like response times).