Page 1 of 2
188000 + syslog entries a sec
Posted: Thu Aug 13, 2015 3:23 pm
by benhank
And still running strong, no issues.
vector-good-job-symbol-29548614.jpg
Re: 188000 + syslog entries a sec
Posted: Thu Aug 13, 2015 3:25 pm
by jolson
Happy to hear it!

Re: 188000 + syslog entries a sec
Posted: Thu Aug 13, 2015 3:30 pm
by tmcdonald
188,000 reasons per second to use Nagios :)
Re: 188000 + syslog entries a sec
Posted: Fri Aug 14, 2015 8:57 am
by WillemDH
Wow, I'm not even at 1000 / sec. But I just got confirmation I get 500 GB extra storage. so more is coming... :p Can i ask how many NLS nodes you have, how many storage they are consuming? retention period? Virtual servers? Or physical? How many cpu's? (Plus minus)
Tx!

Re: 188000 + syslog entries a sec
Posted: Fri Aug 14, 2015 9:04 am
by scottwilkerson
Nice!
@benhank - Can you share how many instances in your cluster and what type of hardware you are running on?
Great story...
Re: 188000 + syslog entries a sec
Posted: Mon Aug 17, 2015 9:57 am
by benhank
Code: Select all
221,444,583Documents
90.3GBPrimary Size
181.5GBTotal Size
2Data Instances
2972Total Shards
298Indices
Code: Select all
2Total Instances
0Client
2Master/Data
10Processors
0%Process CPU
36.12 GBMemory Used
154.62 MBSwap
2,368.26 GBTotal Storage
2,172.45 GBFree Storage
469.13 GBData Read
5,498.76 GBData Written
5,967.89 GBI/O Size
Code: Select all
Cluster Health
Status Green
Timed Out? false
# Instances 2
# Data Instances 2
Active Primary Shards 1486
Active Shards 2972
Relocating Shards 0
Initializing Shards 0
Unassigned Shards 0
Re: 188000 + syslog entries a sec
Posted: Mon Aug 17, 2015 11:31 am
by scottwilkerson
@benhank - I'm a little confused on your math. Was the 188k syslog entries a sec just a quick burst/spike?
What you posted only shows 221m documents indexed, which would be only ~ 19 hours of logs @ 188k/sec.
Re: 188000 + syslog entries a sec
Posted: Mon Aug 17, 2015 11:33 am
by benhank
hey pal I am a sysadmin not a mathamatician... It's the thought that counts =D
Re: 188000 + syslog entries a sec
Posted: Mon Aug 17, 2015 4:11 pm
by tmcdonald
188k/s seems a little high, but for a 2-instance cluster it looks to be humming along quite nicely.
Now I need to do that manager thing and request that we go back to being boring and close up this thread. That cool?
Re: 188000 + syslog entries a sec
Posted: Tue Aug 18, 2015 10:42 am
by benhank
yeah, but i want to apologize for the bad math. My main point was that we have a lot of stuff being processed by NLS, and it is working fine.
You guys made a product that delivers on its stated capabilities.
I spent months trying to get the ELK stack working. NLS was an easy install,as promised, and it makes setting up remote devices simple.
Kudos and good job to the team.
Ben