Change Details

Since the move to k8s, thumbor logging has not been logstash-compliant. We need to build succinct log messages about requests and errors and return them via stdout. Additionally, we need to ensure that we log `client-ip` in our http request messages, as tornado's logger is currently logging incorrect IP addresses (currently kubernetes node IPs).

Since the move to k8s, thumbor logging has not been logstash-compliant. We need to build succinct log messages about requests and errors and return them via stdout. Additionally, we need to ensure that we log `client-ip` in our http request messages, as tornado's logger is currently logging incorrect IP addresses (currently kubernetes node IPs). Other problems (requests by DJ): - messages are currently in the "log" field and not normalised - the level is inside the log line and not the level field - It would be very helpful if we can put the name of the file being converted into a separate field, indexed, so that you can filter on the file extension. - It would be very helpful to have the thumbor plugin handling the actual conversion process to be in a separate field.

Since the move to k8s, thumbor logging has not been logstash-compliant. We need to build succinct log messages about requests and errors and return them via stdout. Additionally, we need to ensure that we log `client-ip` in our http request messages, as tornado's logger is currently logging incorrect IP addresses (currently kubernetes node IPs). Other problems (requests by DJ): - messages are currently in the "log" field and not normalised - the level is inside the log line and not the level field - It would be very helpful if we can put the name of the file being converted into a separate field, indexed, so that you can filter on the file extension. - It would be very helpful to have the thumbor plugin handling the actual conversion process to be in a separate field.