Mirror/zeek
1
0
Fork 0
mirror of https://github.com/zeek/zeek.git synced 2025-10-13 03:58:20 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 2

Move DataSeries and ElasticSearch into plugins.

Browse source
This commit is contained in:
Robin Sommer 2014-08-06 18:43:11 -07:00
parent 8031da4ee7
commit 8737eae906
35 changed files with 9 additions and 2297 deletions
Download patch file Download diff file Expand all files Collapse all files

2
scripts/base/frameworks/logging/__load__.bro
View file

@ -1,7 +1,5 @@
@load ./main
@load ./postprocessors
@load ./writers/ascii
@load ./writers/dataseries
@load ./writers/sqlite
@load ./writers/elasticsearch
@load ./writers/none

60
scripts/base/frameworks/logging/writers/dataseries.bro
View file

@ -1,60 +0,0 @@
##! Interface for the DataSeries log writer.
module LogDataSeries;
export {
## Compression to use with the DS output file. Options are:
##
## 'none' -- No compression.
## 'lzf' -- LZF compression (very quick, but leads to larger output files).
## 'lzo' -- LZO compression (very fast decompression times).
## 'zlib' -- GZIP compression (slower than LZF, but also produces smaller output).
## 'bz2' -- BZIP2 compression (slower than GZIP, but also produces smaller output).
const compression = "zlib" &redef;
## The extent buffer size.
## Larger values here lead to better compression and more efficient writes,
## but also increase the lag between the time events are received and
## the time they are actually written to disk.
const extent_size = 65536 &redef;
## Should we dump the XML schema we use for this DS file to disk?
## If yes, the XML schema shares the name of the logfile, but has
## an XML ending.
const dump_schema = F &redef;
## How many threads should DataSeries spawn to perform compression?
## Note that this dictates the number of threads per log stream. If
## you're using a lot of streams, you may want to keep this number
## relatively small.
##
## Default value is 1, which will spawn one thread / stream.
##
## Maximum is 128, minimum is 1.
const num_threads = 1 &redef;
## Should time be stored as an integer or a double?
## Storing time as a double leads to possible precision issues and
## can (significantly) increase the size of the resulting DS log.
## That said, timestamps stored in double form are consistent
## with the rest of Bro, including the standard ASCII log. Hence, we
## use them by default.
const use_integer_for_time = F &redef;
}
# Default function to postprocess a rotated DataSeries log file. It moves the
# rotated file to a new name that includes a timestamp with the opening time,
# and then runs the writer's default postprocessor command on it.
function default_rotation_postprocessor_func(info: Log::RotationInfo) : bool
{
# Move file to name including both opening and closing time.
local dst = fmt("%s.%s.ds", info$path,
strftime(Log::default_rotation_date_format, info$open));
system(fmt("/bin/mv %s %s", info$fname, dst));
# Run default postprocessor.
return Log::run_rotation_postprocessor_cmd(info, dst);
}
redef Log::default_rotation_postprocessors += { [Log::WRITER_DATASERIES] = default_rotation_postprocessor_func };

48
scripts/base/frameworks/logging/writers/elasticsearch.bro
View file

@ -1,48 +0,0 @@
##! Log writer for sending logs to an ElasticSearch server.
##!
##! Note: This module is in testing and is not yet considered stable!
##!
##! There is one known memory issue. If your elasticsearch server is
##! running slowly and taking too long to return from bulk insert
##! requests, the message queue to the writer thread will continue
##! growing larger and larger giving the appearance of a memory leak.
module LogElasticSearch;
export {
## Name of the ES cluster.
const cluster_name = "elasticsearch" &redef;
## ES server.
const server_host = "127.0.0.1" &redef;
## ES port.
const server_port = 9200 &redef;
## Name of the ES index.
const index_prefix = "bro" &redef;
## The ES type prefix comes before the name of the related log.
## e.g. prefix = "bro\_" would create types of bro_dns, bro_software, etc.
const type_prefix = "" &redef;
## The time before an ElasticSearch transfer will timeout. Note that
## the fractional part of the timeout will be ignored. In particular,
## time specifications less than a second result in a timeout value of
## 0, which means "no timeout."
const transfer_timeout = 2secs;
## The batch size is the number of messages that will be queued up before
## they are sent to be bulk indexed.
const max_batch_size = 1000 &redef;
## The maximum amount of wall-clock time that is allowed to pass without
## finishing a bulk log send. This represents the maximum delay you
## would like to have with your logs before they are sent to ElasticSearch.
const max_batch_interval = 1min &redef;
## The maximum byte size for a buffered JSON string to send to the bulk
## insert API.
const max_byte_size = 1024 * 1024 &redef;
}