PaStasH (pastaʃ'ʃ-utta) is a NodeJS multi I/O processor supporting ingestion, decoding, interpolation and correlation of data - be it logs, packets, events and beyond. PaStash supports the Logstash configuration format and delivers cross-functionality comparable to "Beats" with custom modules, providing a flexible and agnostig data pipelining tool.
paStash is a tool to manage spaghetti I/O with input, processors and output modules for all seasons and protocols.
- lower memory footprint
- lower cpu footprint
- faster startup delay
Moreover it's written in NodeJS, which is a perfect language for programs with many IO.
paStash configuration is compatible with logstash. You can replace a logstash node by a paStash one in most cases. The data are formatted in the same way to be compatible with logstash UIs.
The architecture is identical to logstash architecture. You have to instanciates plugins with the paStash core. There are three type of modules:
- inputs plugins: where datas come into paStash. Examples: file, zeromq transport layer
- filter plugins: extract and manipulate fields from logs, like timestamps. Example: regex plugin
- outputs plugins: where datas leave from paStash: Examples: ElasticSearch , zeromq transport layer.
A typical paStash deployement contains agents to crawl logs and a log server.
On agent, paStash is configured whith inputs plugins to get logs from your software stack, and one output plugin to send logs to log server (eg. zeromq output plugin).
On log server, logs come trough a zeromq input plugin, are processed (fields and timestamps extraction), and send to ElasticSearch.
sudo npm install -g pastash
- Install NodeJS, version >= 0.12
- Install build tools
- Debian based system:
apt-get install build-essential
- Centos system:
yum install gcc gcc-c++ make
- Debian based system:
- Install zmq dev libraries: This is required to build the node zeromq module.
- Debian based system:
apt-get install libzmq1
. Under recent releases, this package is present in default repositories. On ubuntu lucid, use this ppa. On debian squeeze, use backports. - Centos 6:
yum install zeromq zeromq-devel
. Before, you have to add the rpm zeromq repo :curl http://download.opensuse.org/repositories/home:/fengshuo:/zeromq/CentOS_CentOS-6/home:fengshuo:zeromq.repo > /etc/yum.repos.d/zeromq.repo
- Debian based system:
- Clone repository:
git clone git://github.com/sipcapture/pastash.git && cd pastash
- Install dependencies:
npm install
.
The executable is bin/pastash
There are two format for configuration. The legacy format use urls. The new one is identical to the logstash config format.
Note : multiple configuration files can be used in parallel with the --config_dir
switch.
Example for an input file
input {
file {
path => '/tmp/toto.log'
}
}
You can use if
to have an event dependent configuration. See here for details.
As for urls, config can be specified
- directly on the command line
- in a file (use the
--config_file
switch) - in all files in a directory (use the
--config_dir
switch)
Note : the implementation is young, all bugs reports are welcome. Note : both formats can be mixed.
--log_level
to change the log level (emergency, alert, critical, error, warning, notice, info, debug)--log_file
to redirect log to a log file.--patterns_directories
to add some directories (separated by ,), for loading config for regex plugin and grok plugins. Grok patterns files must be located under agrok
subdirectory for each specified directory.--db_file
to specify the file to use as database for file inputs (see below)--http_max_sockets
to specify the max sockets of http.globalAgent.maxSockets. Default to 100.--alarm_file
to specify a file which will be created if paStash goes in alarm mode.
Config file for an agent:
input {
file {
path => "/var/log/nginx/access.log"
}
}
output {
zeromq {
address => ["tcp://log_server:5555"]
}
}
Config file for log server:
input {
zeromq {
address => ["tcp://0.0.0.0:5555"]
}
}
filter {
regex {
pattern => http_combined
}
}
output {
elasticsearch {
host => localhost
port => 9200
}
}
See our wiki for many more examples
You can add easily add your plugins :
Manually :
- create a directory layout on the path of your choice :
/var/my_plugins/inputs
,/var/my_plugins/outputs
,/var/my_plugins/filters
- set the NODE_PATH variable to
NODE_PATH=/var/my_plugins:/node_logstash_path/lib
- add your plugins in
inputs
,outputs
orfilters
directory. In the plugin code, you can reference base plugins withvar base_filter = require('lib/base_filter');
- reference your plugin as usual.
With native packaging
The plugins must be deployed in /var/db/pastash/custom_plugins
. All subdirectories already exists. The NODE_PATH is already set.
- USR1: stoping or starting all inputs plugins. Can be used to close input when output targer are failing
- USR2: see below file output plugin
- File
- Syslog
- ZeroMQ
- Redis
- HTTP
- Websocket
- TCP / TLS
- Google app engine
- AMQP
- MQTT
- SQS
- NetFlow
- sFlow
- Bencode
- Freeswitch ESL
- Asterisk AMI
Common concepts / parameters :
- Regex
- Grok
- Mutate Replace
- Grep
- Reverse DNS
- Compute field
- Compute hash
- Compute date field
- Split
- Truncate
- Rename
- Multiline
- Json fields
- Geoip
- Eval
- Bunyan
- IPProto
- HTTP Status Classifier
- Remove field when equal
- Mustache
Common concepts / parameters :
Apps with embedded parsers :
Common concepts / parameters :
paStash Copyright 2016 - 2017 QXIP BV
node-logstash Copyright 2012 - 2014 Bertrand Paquet
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.