We're using elasticsearch to log requests and responses. We're also running extra analysis to do things like add an outlier score. Similar to the example in this rep but in our case we'll actually enrich all the documents with an extra attribute so that we can graph that. We're also looking at graphing other features in the documents (request and response variable values), so kind of using elastic as a time-series database but not like prometheus metrics as the source data is all the individual requests as documents. I guess this is an unusual use-case so presumably there are no examples close to it. Just wondering if anyone has any observations that might help us. Happy to raise this somewhere else if preferred.
We're using elasticsearch to log requests and responses. We're also running extra analysis to do things like add an outlier score. Similar to the example in this rep but in our case we'll actually enrich all the documents with an extra attribute so that we can graph that. We're also looking at graphing other features in the documents (request and response variable values), so kind of using elastic as a time-series database but not like prometheus metrics as the source data is all the individual requests as documents. I guess this is an unusual use-case so presumably there are no examples close to it. Just wondering if anyone has any observations that might help us. Happy to raise this somewhere else if preferred.