Open dusts66 opened 6 years ago
@dusts66
What does your config/kafka.yml
look like?
Here is the config @sghaskell
hec:
# Splunk HEC host or VIP to load balanced HEC instances
host: 10.26.155.169
# HEC port - default 8088
port: 8088
# App specific UUID used by HEC - generate your own using 'uuidgen' in Linux
#channel: 408c8da1-b9b1-460b-aeb4-56426865bca1
channel: fbf00d67-0287-49b3-8cf7-54db7ee6f2d1
# HEC token for your input
token: 0688D98D-06FD-4AA9-B195-19EAE1D89FAE
# Set input sourcetype
sourcetype: cisco:cmx
# Set input source
source: ITAM_Splunk
# Use HTTPS
# Valid values: True (HTTPS)|False (HTTP)
use_https: False
# Verify SSL certificate
# Valid values: True|False
verify_ssl: False
# use gzip compression when sending data to HEC
# Valid values: True|False
use_compression: False
# Compression level 0-9; 0=none, 1=fastest/least, 9=slowest/most (default: 9)
compresslevel: 5
kafka:
# List of Kafka brokers <host>:<port>
brokers:
- xrdcldbda010001.unix.medcity.net:9092
- xrdcldbda010002.unix.medcity.net:9092
- xrdcldbda010003.unix.medcity.net:9092
# Zookeeper hostname/ip
zookeeper_server: xrdcldbdn010002.unix.medcity.net
# Zookeeper port
zookeeper_port: 2181
# Arbitrary group name used for coordinating workers in topic
consumer_group: shark
# Kafka topic name
topic: asset_mang
# Initial offset - Set to earliest for earliest offset or latest for latest offset
# Valued values: earliest|latest
initial_offset: latest
# If you have librdkafka installed and pykafka was built against it, use this to increase speed
# See: https://github.com/Parsely/pykafka#using-the-librdkafka-extension
# Valid values: True|False
use_rdkafka: False
general:
# Set to number of partitions in the topic for maximum parallelism and throughput
# Do not exceed number of cores on your system - spread across multiple machines
# Do not set workers across all instances > # of partitions in topic - will result in idle wokrers
# Set to 'auto' for python multiprocessing to figure out core count
# Valid values: auto|<number_of_cores>
workers: 3
# Number of messages to send to Splunk in each HTTP call to HEC
# Data will not be sent until there is batch_size events in the topic.
# Increase or decrease depending on activity in topic.
# WARNING: If you set this to 1 it will increase the activity on your network and decrease
# throughput to the HEC.
batch_size: 1024
network:
# Backoff policy settings
# Number of retry attempts before quitting
retry_attempts: 5
# Sleeptime between retries (seconds)
sleeptime: 60
# maximum sleeptime (seconds)
max_sleeptime: 300
# sleep time (seconds) multiplier applied to each iteration
sleepscale: 1.5
# random jitter (seconds) introduced to each iteration, random between [-jitter, +jitter]
jitter: 5
logging:
# Valid values: debug|info|warning|error|critical
loglevel: info
It looks like you're missing the entire ssl:
section from your config here. Add that back in and modify it to use_ssl: False
, if you're cluster isn't secured with SSL. It is expecting that section to be present in the config.
I'll add some checks to ensure required key:values are present in the config and warn the user with better errors. Thanks for reporting this.
Starting the consumer errors out with the following error:
Not sure what this is about