4.0.0
Well a lot a rewrite into this release.
New feature / Improvement
- Buffering Same as loggregator itself, we use diodes (ring buffering) so now
Firehose message are bufferize before being send to syslog, buffer size could be parametize--logs-buffer-size=10000 - Draining Due to the buffering we now drain the remaining message when the app being stop
Draining will timeout after 30s (not configurable) - Retrying connection to Firehose in case of deconnexion
- Embed Stats Server
Provide information about buffer size and so on at http://IP:8080/stats
Means you can use http healthcheck with CF
Could be enable by using--enable-stats-server
Still in early phase - Unwanted event are now earlier filtered out means the buffer only contain wanted event.
Know issue
- Yes some test are missing ! Need more time to add them
- Race(s) condition in the Stats Server no real impact but still
- Drain timeout is not configurable
- Stats Server port is not configurable