Creates ELK cluster and watches for logs directory.
So far parses only basic supportconfig, syslog from messages*.txt
and Ceph logs. For other files more filters needed to be written.
Please check .conf
files for excluded files and directories.
- setup docker, and docker-compose
- check that your system has correct parameters for elk in docker
sysctl -w vm.max_map_count=262144
- clone this repo,
cd elk_supportconfig
docker-compose up
- unpack your supportconfig (or supportconfigs if you have cluster) to logs directory
- wait untill all events are imported (there is a simple event monitoring every second is printing to stdout/logs)
- go to Kibana interface in web browser http://your_host:5601 (usually localhost)
- go to Dashboard->Ceph general dashboard
- don't forget to change Time Filter in the right corner to something more sane (defaults to 15 Min) - maybe to "Year to date"
- use filters, searches and Time Range to filter and look for you data
- Cltl+C to destroy cluster and
docker-compose rm -v -f
to delete container's data
You can create your own searches in Discover. After saving your search/dashboard you can import it from .kibana index and place under kibana directory. You need to delete
To import index/search/dashboard:
- create new index pattern
.kibana*
in Kibana (don't use Time Filter) - go to Discover, choose
.kibana*
pattern, find your object, go to JSON and copy paste it to file - you need to remove
"_source": {
and corresponding}
without removing data inside - you can use one object per json file in kibana directory
- or use
curl
to query elasticsearch directly for .kibana index
- create Beat to import
ceph report
json - create script or reuse beat to import/export indices, searches, dashboards
- more parsers
- better indices and fields