|
Data Graham posted:Or plugs. It’s just so convenient though One of my friends calls cables/ports/whatever with the terms insertive/receptive.
|
#
?
Jul 24, 2020 18:43
|
|
|
|
| # ? May 26, 2024 19:52 |
|
|
Methanar posted:I'm still waiting for all plumbing components to have their gendered names removed. Apparently this was quite the problem when designing the ISS as neither USA nor Russia wanted to be the female half.
|
|
#
?
Jul 24, 2020 18:44
|
|
|
Arquinsiel posted:Apparently this was quite the problem when designing the ISS as neither USA nor Russia wanted to be the female half. I feel like it would be the other way around. The female half of something is usually the main thing the other thing plugs into. Being the female half would imply you're the ACTUAL ISS and the other piece is just an attachment. Then again, people are idiots.
|
|
#
?
Jul 24, 2020 18:48
|
|
|
Sirotan posted:Words matter, and there's no reason to continue using whitelist/blacklist or master/slave as tech terms when substitutes without racial overtones exist other than "well that's the way it's always been!" Is master/slave even a term used anymore? I am pretty sure windows removed all the terminology long ago. I certainly haven't heard anyone use that term since... the 90's?
|
|
#
?
Jul 24, 2020 19:54
|
|
|
Sickening posted:Is master/slave even a term used anymore? I am pretty sure windows removed all the terminology long ago. I certainly haven't heard anyone use that term since... the 90's? It's still used in the context of database replication, and I renamed an abstract class called "master" just yesterday.
|
|
#
?
Jul 24, 2020 19:55
|
|
|
"Golden Master" for CD/DVD disks ?
|
|
#
?
Jul 24, 2020 19:59
|
|
|
whitelist/blacklist were still pretty heavily used all over the place, but master/slave hasn't been used in hardware terms in a long time, though yeah database world it was still a thing.
|
|
#
?
Jul 24, 2020 20:00
|
|
|
orange juche posted:master/slave hasn't been used in hardware terms in a long time Which has little to do with it being consciously phased out in place of other terminology and everything to do with SATA basically replacing IDE entirely at this point.
|
|
#
?
Jul 24, 2020 21:28
|
|
|
In stacking switches, I've generally seen the primary member of the stack still called the "master." Of course, even that sentence shows how easy that terminology would be to replace: primary, main, controlling, lots of options.
|
|
#
?
Jul 24, 2020 22:06
|
|
|
replace "master recording" with "the good one. no, the GOOD one. with the saxophone"
|
|
#
?
Jul 24, 2020 22:09
|
|
|
Weedle posted:replace "master recording" with "the good one. no, the GOOD one. with the saxophone" +1 except with "more snare crack"
|
|
#
?
Jul 25, 2020 00:52
|
|
|
I know with casting you make a "master" mould, but there's no slave moulds, you just make "copies".
|
|
#
?
Jul 25, 2020 03:43
|
|
|
modern society has already come up with an acceptable substitute term and it's 'daddy.' posting in SH/SC heyoooooooo
|
|
#
?
Jul 25, 2020 03:50
|
|
|
 Now fix this one.
|
|
#
?
Jul 25, 2020 03:53
|
|
|
Strumpie posted:modern society has already come up with an acceptable substitute term and it's 'daddy.' Brb renaming my databases Daddy and littles
|
|
#
?
Jul 25, 2020 03:56
|
|
|
king(queen) / simp
|
|
#
?
Jul 25, 2020 03:58
|
|
|
Geemer posted:
|
|
#
?
Jul 25, 2020 04:03
|
|
|
Weedle posted:replace "master recording" with "the good one. no, the GOOD one. with the saxophone" _final_FINAL__[use this one] (3).docx
|
|
#
?
Jul 25, 2020 04:32
|
|
|
The correct terms are dom and sub, obviously.
|
|
#
?
Jul 25, 2020 09:21
|
|
|
Bourgeois/Proletarian?
|
|
#
?
Jul 25, 2020 13:17
|
|
|
InfoSec: were gonna move splunk to AWS We currently ingest 2-3TB/day
|
|
#
?
Jul 25, 2020 15:31
|
|
|
AlexDeGruven posted:InfoSec: were gonna move splunk to AWS Last I heard (May) it was well over 4TB. I don't trust our infosec team to build anything at all. They can't even look at the current logging infrastructure without it falling over. Which is fantastic, because our internal library kills the application processes if it can't contact splunk!
|
|
#
?
Jul 25, 2020 15:44
|
|
|
AlexDeGruven posted:InfoSec: were gonna move splunk to AWS That definitely looks like something that is going to be expensive, but I feel like if you are ingesting 2-3tb of things a day you might have some other issues.
|
|
#
?
Jul 25, 2020 16:06
|
|
|
Motronic posted:Which has little to do with it being consciously phased out in place of other terminology and everything to do with SATA basically replacing IDE entirely at this point. It may have more than you think, since this isn't the first time I remember the industry having this conversation. There is significantly less scoffing this time, tho.
|
|
#
?
Jul 25, 2020 16:17
|
|
|
Sickening posted:That definitely looks like something that is going to be expensive, but I feel like if you are ingesting 2-3tb of things a day you might have some other issues. Part of it is being in the financial industry. Part of it is just volume of systems. A big part of it is our lolinfosec team.
|
|
#
?
Jul 25, 2020 16:49
|
|
|
xsf421 posted:Last I heard (May) it was well over 4TB. I don't trust our infosec team to build anything at all. They can't even look at the current logging infrastructure without it falling over. Which is fantastic, because our internal library kills the application processes if it can't contact splunk! That's a catastrophe waiting to happen.
|
|
#
?
Jul 25, 2020 17:41
|
|
|
Yeah make sure you keep us updated on it's inevitable implosion
|
|
#
?
Jul 25, 2020 17:44
|
|
|
I migrated Splunk for a large healthcare that ingests 2TB/day to Azure. Annual costs come out to about $800k. Currently working on adding additional log sources that'll bring ingest up to their current Splunk license of 5TB/day, they'll probably be spending 1.2-1.5 a year just on Azure infrastructure.
|
|
#
?
Jul 25, 2020 17:57
|
|
|
spankmeister posted:That's a catastrophe waiting to happen. Don't worry, we have the finest people with infosec bachelor's degrees and no actual engineering experience building out this super critical infrastructure!
|
|
#
?
Jul 25, 2020 20:37
|
|
|
Sickening posted:That definitely looks like something that is going to be expensive, but I feel like if you are ingesting 2-3tb of things a day you might have some other issues. I know video games that do multiples of this a day.
|
|
#
?
Jul 25, 2020 23:00
|
|
|
Biowarfare posted:I know video games that do multiples of this a day. Ingesting large amounts of logs in itself isn't bad. I would think that the video game services/infrastructure that create those kinds numbers were also designed to. When an average org is throwing out big daily numbers like that, the first thing I think is that throwing every source they can at it and aren't curating what is being sent. Every org I have been a part of has burned cash and lazy splunk implimentations in some form or another.
|
|
#
?
Jul 26, 2020 00:58
|
|
|
I've spent the few months pushing back and asking "Why are we logging and alerting on this? Will this produce anything actionable?" and the answer has always been "Well, we can log it, so of course we should. And you never know when you're going to need to do something about... something." Alarm fatigue is a real thing, but the bosses who don't have to actually do anything about it refuse to realize it.
|
|
#
?
Jul 26, 2020 02:05
|
|
|
My boss is really good about that, only alerts that we need to be woken up by or even truly care about during the day go into Opsgenie. It’s an iterative process, any time we run across a situation where a holy poo poo alarm should have happened, we get the alerting in place to cover the next time.
|
|
#
?
Jul 26, 2020 02:28
|
|
|
That's the best way to do it if you can afford* to have poo poo fall over before you catch it. The "log everything" method is... not smart. *ETA: "afford" defined as "won't have a regulator up your rear end killing the business if you get popped" here.
|
|
#
?
Jul 26, 2020 02:42
|
|
|
Wizard of the Deep posted:I've spent the few months pushing back and asking "Why are we logging and alerting on this? Will this produce anything actionable?" and the answer has always been "Well, we can log it, so of course we should. And you never know when you're going to need to do something about... something." The best policy I've ever seen for an alerting system was to forbid email alerts. It's either a formal incident with an SLA and all that jazz, or it's just a meter on a dashboard. If it's not important enough to make someone do a bunch of paperwork every time that it pops up, it isn't an alarm. This means that if something goes off five times a day, someone gets pissed and starts screaming that all they do all day is clear incidents because CPU went over 50% and either they fix the threshold, fix the cpu problem, or turn off the cpu alarm.
|
|
#
?
Jul 26, 2020 02:53
|
|
|
We've been actively reducing our Nagios alerts over the last few years, with efforts significantly ramped up over the last couple months. In some cases, we're replacing largely useless checks (swap usage, for example, which by itself is meaningless most of the time) with better ones (heavy active swapping/thrashing, in that case), adding logfile check exclusions, and creating a new check that e-mails the application administrators for certain logfile errors without creating an alert. Stuff like that. In the last 2 months, we reduced our number of tickets generated by alerts by something like 80 percent and we've still got a bunch of stuff in the pipeline. My goal is for any alert that generates a ticket or NOC call to require some sort of action. We're decent at the latter now, but the former still has a way to go.
|
|
#
?
Jul 26, 2020 03:25
|
|
|
I literally made an alert that includes in its message body 'No action is required' 2 weeks ago because my boss wanted one 
|
|
#
?
Jul 26, 2020 03:27
|
|
|
Methanar posted:I literally made an alert that includes in its message body 'No action is required' 2 weeks ago because my boss wanted one We effectively have those as well in certain situations by virtue of utilizing an existing nagios check to notify when a thing happens that we want to know about rather than creating something else that would do it. I don't count something like that exactly where it's just leveraging an existing system rather than creating something new. Used sparingly that's not a huge deal.
|
|
#
?
Jul 26, 2020 03:29
|
|
|
We have a particular alert that goes to our service desk that occurs maybe 2-3 times a month. The email has instructions for handling that include "create critical (P1) incident" and "call <list of names>". So of course, every time one comes in the entire 80-person IT department gets blasted by pages because a P1 was opened. The manager of the team who owns the system is always complaining that the service desk doesn't need to do this, these aren't high priority, why are they calling us, etc. You wrote the loving alert verbiage, Jerry, so either change it or shut the gently caress up. Of course, these definitely aren't P1 incidents, but if we don't do it then we'll just eat poo poo for not following instructions.
|
|
#
?
Jul 26, 2020 04:54
|
|
|
|
| # ? May 26, 2024 19:52 |
|
|
We have SolarWinds everything. I’m the only one who uses it. Any alerts that exist is because I set them up. They go to my mailbox. Shrug.
|
|
#
?
Jul 26, 2020 06:26
|
|