Splunk to cut workforce by 7% after cisco deal | Hacker News Reader

Splunk to cut workforce by 7% after cisco deal

bloomberg.com

167 pointsmfiguiere148 comments

Cisco buying your company should be considered a worst case scenario for most employees. The important thing to remember is not to trust a word they say.

I was at Kenna Security when Cisco bought them (I was only there for a month too, so Kenna hired me knowing the purchase was going to happen). They lied through their teeth about the process. They promised every resume would be reviewed for leveling, but then gave out the worst offers I've ever seen. They were literally dropping people by multiple levels, and if anyone complained they told them they couldn't make personal exceptions. They wouldn't even listen to people's managers about it, it was just a bunch of anonymous resume reviews.

At the same time Kenna leadership said they'd pay for Cobra for me if I didn't want to stay. The VP of Eng, David La France, at Kenna promised me that to my face. Then once they sale went through he refused to talk to me and told me to cover it myself. Ultimately they did back down, and were "kind" enough to give me a whole two weeks severance. However it made me realize that Cisco views it's acquisitions in the worst way possible, and if I ever work at a company that Cisco purchases I will immediately start looking for a new job.

It's worth noting that I've kept in touch with a lot of the Kenna people who stayed (at least initially), and according to my old manager more than half of the people who stayed on are gone now.

> The VP of Eng at Kenna promised me that to my face

The real lesson to learn here is twofold:

- People recruiting / hiring are incentivized to say whatever is needed to close the deal. With the exception of a bold faced lie (I don't think this was that).

- People at the company being acquired are more often than not going to lose most of their power. Including the C-suite. No one is safe and as an employee/victim you won't know the actual arrangements of the deal until the dust settles.

Three fold:

- People forget and circumstances change. Signed documents don't.

Exactly. Get all promises in writing (ala, a contract of sorts). The company certainly would when they get a promise out of you.

Yea, I think more often it is the second point more than the first one. You would be surprised how often makes leaders into liars. Honestly, I would bet that the VP probably had assurances from his leadership that what he was telling you was true.

Sorry to hear that happened, that sucks.

This isn’t a Cisco thing - it’s M&A in general. The sellers need to maintain the appearance of growth before the sale, and the buyers want to rationalise costs. Once they have the IP, or have successfully eliminated the competition through acquisition, those in the old company are really just numbers in an hr system.

When I was at Malwarebytes (I left in 2014) I helped with two acquisitions. In both cases we took care of all of the people in the new company and rolled them into Malwarebytes.

M&A can be done without being ethically and morally bankrupt. It's totally possible. Cisco just doesn't think that's a priority.

Companies, regardless of what the Supreme Court says, are not people; they don't have ethics and morals. All that matters is the balance sheet, the stock price, and the executive bonus.

You can have people running the company change what the company does to be more ethical or moral, but it's like training an elephant to do tricks: an unnatural edge case. Their natural incentives are more like swarms of locusts or jellyfish; the more you feed them, the larger they grow, the bigger impact they have on their environment. No real thoughts, just an urge to consume and expand.

Ethics an morals have financial upside, for people and groups of people (companies). For instance, do to their past behavior, and its publicity, Cisco is likely to lose many more competent employees during an acquisition than had they behaved otherwise.

I would say that a lack of ethics has more financial upside

There are probably several local maxima on the distribution curve. At some point having dogmatic, extreme ethics probably hampers getting everything down, and the worst, most cartoonishly evil company also is leaving money on the table. In the less extreme, I think you probably have three schelling points. 1) High-Integrity companies that promote and use the culture of ethics to attract customers and employees. It probably is industry specific if this works. 2) You have "ethics agnostic", but legal" companies that just do focus on near-term profit maximizations and while they may have individuals trying to make ethical choices, the organizational structure won't reward that outside of the whatever profit it generates. Non ethical behavior that gets bad press or crosses idealogical lines, gets dropped fast if not profitable and arbitrarily depending on leadership otherwise. 3) Then you have companies covering up illegal/fraudulent issues, where the incentives are to double-down and reinforce the behavior.

Depends the environment you're working in - there are several cultures within the US where I think your statement is quite true... but in general people find that working for a company that doesn't care for them is a detriment and will demand higher compensation due to the obvious risk. That all said - you should always be defensive as an employee about how much risk you accept on yourself since the employer/employee relationship can be terminated suddenly and without any real recourse in most circumstances.

Depends on the circumstances. If you are trying to make a valuable product, good ethics will probably serve you better than bad ones in a majority of cases. If you aren't concerned with generating real value and are fine with running your company and reputation into the ground, then sure.

That's an amazing cop out. Companies are run by people, people make the decisions, and ethics/morality can have an impact on the bottom line. Using "it's a company, ethics aren't real" is just an excuse people who make decisions use to act unethically.

> an excuse people who make decisions use to act unethically.

at every level, the incentive is to maximize the profit margins. If acting ethically doesn't increase the profits, then any personal sacrifice on the part of the decision maker to act ethically is only going to get punished (may be not immediately, but certainly some time in the future).

If behaving ethically didn't cost anything, everyone would do it.

> at every level, the incentive is to maximize the profit margins

This is wildly oversimplifying. For example, it theoretically rules out principle-agent problems.

While true, you can somewhat assume that the principle could have already incentivized the agent to align by assigning rewards for increasing profits (aka, bonuses, stock grants based on performance targets etc).

> you can somewhat assume that the principle could have already incentivized the agent to align by assigning rewards for increasing profits

No, you can’t. This is the problem of corporate management. Management will tend to act to maximise profit margins or even the stock price is a bad explanatory model.

>at every level, the incentive is to maximize the profit margins

Over what time frame?

You're absolutely right; shame on whoever made the decisions! In this case, it was certainly within their power to make a better choice, and they should have done.

However, if you want tech jobs to have a higher and guaranteed severance, national legislation might be more effective.

After all, many acquisitions and layoffs are at companies that are in dire financial straits, which limits discretionary spending.

I'm certainly a fan of legislation to require severance. That said I'm also all about shaming bad actors so that other workers are aware of what they can expect when engaging them.

There is no reason to involve businesses in the equation. If your goal is to ensure people have cash in the event they are terminated from their job, then have the government give them cash in the event they are terminated from their job.

Why should the taxpayers bear this cost instead of the business who is choosing to terminate the person, especially in situations like those discussed where it has nothing to do with the performance of the individual?

The government can collect taxes from the businesses, and call it something like unemployment insurance premiums in the US.

There is no reason to over complicate things and require businesses to directly pay, figure out how much cash to always have to be able to pay, hire auditors to make sure they have enough cash, then take them to court when they don’t have enough cash to pay, yadda yadda.

Business sells product or service. Employee sells labor to business. Government takes care of the safety nets.

> That's an amazing cop out. Companies are run by people, people make the decisions, and ethics/morality can have an impact on the bottom line. Using "it's a company, ethics aren't real" is just an excuse people who make decisions use to act unethically.

FWIW, I didn't read parent post as making an excuse, just describing the reality.

In the corporate world, approximately nobody is rewarded for acting ethically, and as such it's a reasonable assumption that corporations will always trend towards unethical behavior (read: any behavior that pushes costs onto some other entity, while bringing profits to the company), because that can be a competitive advantage.

Acknowledging this is a first step in fixing the problem. We need to incentivize companies to do do the "right thing" - where the "right thing" is something other than simply maximizing profits - because we can't rely on "good people making moral choices" when they are incentivized not to do so.

But how? I can’t know if I’m buying xyz software if 3rd level manager Janet acted ethically, or not. There are many other people involved. Companies always claim to be ethical, in some way that might even be true.

That's the point/trick of the status quo. It's impossible as a "consumer" to make choices that can meaningfully impact anybody's behavior in a system like this.

The only way to achieve a change in behavior in corporations is to change the regulatory framework in which they operate.

In the narrow context of the OP, this doesn't even need to be very heavy handed (unlike say environmental devastation, which is a much bigger challenge). One option would be to pass legislation that makes it harder for megacorps like Cisco to eat smaller companies. Another would be to make layoffs relatively more expensive (require better severance, etc) or to limit them during the M&A phase that Splunk is currently in.

> Companies, regardless of what the Supreme Court says, are not people.

What the USSC said was that voluntary associations of people don’t forfeit the rights of the individuals in the association.

Not challenging peoples capabilities doesn't scale well. Did you ever hire someone with 0 interviews, just because they work for another good company? Then why do that after M&A?

Its not unusual for the force cuts to be done via interviews at the new company

> Then why do that after M&A?

Presumably if you trust the organisation enough to pay for the whole thing, you would have a measure of trust in their ability to hire?

Depends what you’re really paying for. If it’s about the IP, the market access, or a particular individual (seen this myself) then the rest of the org might be dead weight.

You're misreading my statement. My point is that they didn't do interviews. They just gave everyone absolutely crap offers, without actually interviewing or evaluating individuals. Hiring people with 0 interviews just because they worked for the purchased company is exactly what happened, and they did it in such a way that people were grossly under leveled.

It often has to do with the size of the company: the larger you are, the weaker your decisionmaking.

I work for a very large corp. As part of a round of layoffs earlier this year, they cut an entire foreign office, to the man. The decision was made from very high up: VPs were informed after it occurred. For some departments, the losses were small. For others, they were crippling: It was a very important group, doing things nobody else did, and which we couldn't cut.

The end result? The VP realized that the cuts had been extremely unwise, and now almost everyone that worked on that group has US visas, and is working from a new office in the US, with American salaries instead of their far, far lower ones. Not a costs savings, not an improvement in capability... just more expensive, and with projects that got delayed for months, as everyone was out for about 6 months.

For US layoffs, the decisions were not made quite that high, but still high enough that people in the know of salaries and performance were extremely confused about who got cut. Some great, cheap people were cut. Some expensive people that are poor performers by any standard remained. But nobody that managed ICs was involved with the decision making.

A large organization either makes very slow decisions, or acts basically blind. Sometimes they really fail, and do both!

I'd summarize it as: at large scale, M&A cuts happen only at the department budget level.

No one is looking at individual people or teams.

The pain is real. Those layoffs forever changed the company. It killed the culture, ultimately the controlling stock holders were the decision makers. It is their company and they had a message to send.

The message was that they were in control. That they could cut off their nose to prove a point. This layoff event will be their downfall.

I’m considering joining a startup, I think they could be acquired before a year is up by someone, their typical 1 year cliff/4 yr total options wouldn’t be vested. The usual answer is the new company will surely want the engineers and offer them new stock or equity, if not too bad. As is common, half the comp is in options, and lower cash comp at this startup.

I’m considering whether I should ask for my options to immediately vest in the event of a buy out or liquidity event. It’s not that much money or options that the overall company will care. Should I ask for 1 year, or all 4 years, I’m uncertain if it could poison the well. Is this even something they would do?

M&A of stagnant businesses, yes. Cisco, Oracle, and private equity are acquirers to worry about. They buy after growth has stalled and milk users that have high switching costs.

Acquirers of growing businesses generally want that growth to continue and can "optimize" by hiring less.

Worst case scenario? That seems unlikely.

Worst case scenario would be employee stock options and RSU grants remaining or becoming worthless, like in the case of a private company that never gains enough traction to exit or exits at a low valuation.

Usually in a public acquisition, shares are purchased at a price premium.

E.g., many Twitter employees got to cash out at a stock price well above the value of the company.

Every employee should see their role as temporary. This isn’t the lifetime employment pension days anymore. I see many opportunities for employees to walk away as a positive.

> Every employee should see their role as temporary. This isn’t the lifetime employment pension days anymore. I see many opportunities for employees to walk away as a positive.

Good advice in general, but not a hard rule. There are roles where you need to expend a lot of energy learning proprietary stuff (processes, languages, frameworks, business rules). If you're in one of these roles it can give you a ton of leverage, but tread carefully around egos (you could get fired out of spite if you flex too hard, and maybe that's fine, but generally it's lose / lose).

Every employee should see their role as temporary. This isn’t the lifetime employment pension days anymore. I see many opportunities for employees to walk away as a positive.

You're right, of course, with just a nit of misinformation here that always bugs me: there are still lots of places that have pensions, just not places the HN demographic considers worthy of working at. I have one from my time at IBM, one from my telco days, and my current employer (financial) has a pretty generous one.

Yes, I was part of the Kenna acquisition and got my equity paid out. That was nice.

I've also had a company get acquired by private equity. I got nothing for my shares and they shut the team (and the product) down within two years.

For a contrary perspective, this wasn't my experience with being acquired by Cisco. I had my issues with the process, but I felt well-treated throughout and was never lied to.

Agreed. I was also part of the Kenna acquisition after being there for 2 years and am still there.

My equity got paid out quickly, I have since gotten raises and bonuses, and the only shittiness has been all the mega-corp nonsense (omg the trainings) and worse health insurance.

Cisco is not a nightmare acquirer, that would be Amazon. If AMZN had acquired us I'd have left immediately.

Isn't this the usual case for an acquisition by a big company? People are just resource units to them, same as any other asset...

So the Kenna execs lied to you about the Cisco merger and what it means. OK.

FWIW, I've never be bought by Cisco, but my best friend has been at 3 startups bought by Cisco. He never enjoyed working for Cisco, but always said Cisco was very clear about their process, and it basically worked the same way each time, except when execs at his startup threw monkey wretches into things. YMMV.

They weren't clear about anything with me- I had to really push to even know information about what the new insurance would look like. It was a very stressful situation, and while I do blame the horrible leadership at Kenna for a lot of it I can definitely say that Cisco did not help matters at all.

In any M&A situation, you want to start looking for a new job ASAP. They don't _always_ turn out bad, but they often do, and you want to be in a position where you have options.

Clearly you misunderstood them; they were going to send a venomous snake to your house and not even charge you for it.

Interesting. I too was at Kenna, hired a few months before the acquisition was over. Don't remember hearing promises of interviewing. As far as I could tell, my role was leveled based on salary. Since I was an IC and promised shares would covert to cash, it all seemed OK to me. Perhaps I wasn't paying close enough attention?

Regardless, a dream offer landed before my shares vested, so I made the hard choice to leave empty handed. There did appear to be a lot of turnover before and after the acquisition. Though during M&A that's to be expected IME, so experiences may vary.

I didn't say they were going to interview people, just that I was told that each person would be reviewed separately and given custom offers based on their actual background. Then after being strung along for a month on that lie, I was told that they were going to level based on title instead without looking at the individuals background. As a result Cisco vastly under leveled me in their offer (I was able to get another job at the appropriate level pretty trivially, so it's not like I'm over judging my qualifications here).

https://i.kym-cdn.com/photos/images/newsfeed/001/876/233/4c9...

I was at Cisco bought company, what a disaster it became 5 years later. Politics after politics, made me wonder why they even bought it.

I think everyone whose equity became worth actual US dollars disagrees that Cisco buying them is a "worst case scenario". You would not understand that perspective however, since the vesting cliff is longer than 3 weeks.

Sorry that happened to you, I have been through a few acquisitions as both a survivor and a casualty. Being a casualty is certainly no fun. However, if you only have a month of time with your organization and it gets acquired, unless you are some hotshit super hire—you are probably getting RIF’ed, no matter who the acquiring company is and no matter what promises were made to you at any point. Even getting promises in writing doesn’t really help unless they are in your employment contract and most good employment contracts will give the hiring company a legal out for just about any reason within a probationary period (which is often around 90 days).

The two weeks of severance was actually pretty generous too for your time invested.

Typical post deal RIF, but bigger question - what do people think of the "observability" space from here? My sneaking suspicion is that outside vertical integrations like this.. they are going to see some revenue issues in this downturn.

Outside of huge FAANG infra, paid observability seems like a "nice to have" rather than a "must have".. the type of thing you cut before you cut engineering staff, and extend your runway.

You can cobble together something worse with FOSS.. And biggest question is - who cares about p99 latencies anyway when your 2 year runway has shrunk to 6 months, ZIRP is over and you are doubtful of your next funding round.

I can't imagine another whale like Coinbase writing a $65M/year check to a Datadog at this point... "Things you only see at the top"

I get to peak into a few different financial services org's IT strategies (and where infosec overlaps) and in every org I have seen ruthless cutting (hundreds of millions of dollars in one case) if there is any hint of overlap or the ability to go without. Have also seen aggressively playing VARs/reseller off of each other to drive down cost.

I would not want to be someone attempting to GTM in this space right now unless you are very confident in your value prop and your ability to communicate that your customer absolutely cannot live without your product.

Top passed, winter is coming.

Yes, and Splunk is likely the most overpriced offering across all vendors, both in terms of absolute cost and in in terms of how overpriced it is relative to how good it is. So easy to see how it would be on the chopping block.

doubts in datadog

Yes, datadog and newrelic are both crazy expensive. Seems like the space could really use a disruptive competitor with a totally different business model and approach. I think a lot of customers would be OK with higher latency if it could cost 90% less. Would also be cool if you could somehow use your spare on-prem compute to shoulder some of the load to bring the cost to serve down.

in every org I have seen ruthless cutting (hundreds of millions of dollars in one case) if there is any hint of overlap or the ability to go without

Can confirm; I'm on a team doing this analysis right now. And, frankly, it's a really good idea.

When I started asking questions like "why do we have 4 different proxy technologies solving the same use case", "why are we using an on-prem solution over here and a SaaS over there to do the same thing" and "why are you buying this tool when we have an MSA with one of their competitors that will save us 60% on the licensing", it was enlightening. Occasionally, there was a really good reason ("it has feature X which makes us Y percent more efficient"...ok, that makes sense) but almost always we got something like "I didn't know anyone was using Product D when we purchased Product E".

I dunno if it's going to be 100s of millions in savings, but we're definitely going to see 10s of millions.

Have also seen aggressively playing VARs/reseller off of each other to drive down cost.

Hey now...that's the fun part.

It’s hard to get application performance monitoring (APM) tools using FOSS because of the complexity of writing libraries.

You could ideally work most things with FOSS. Use opentelemtry extensions to a time dries DB of your choice. Use elk or opensearch for logs, osquery for Linux, native cloud monitoring tools that come out of the box, etc.

Problem is it takes a lot of setup and maintenance. It can bite you back in the worst of times - when you’re trying to work an outage and find out alerting is dead because your fluentbit daemon is hung and no logs came through.

Or you can spending a day or two creating enough things to track the "end to end" numbers and debug problems the hard way when they appear.

Modern APM is completely out of proportion, and most people that use it do absolutely not get enough benefit for paying for those systems. My impression is this is yet another of those Gartner-created hypes that survive as a meme on the high-management minds and don't survive touching the ground.

Worse yet I saw a lot of APM-or-nothing mentality from CTOs & senior management such that we skipped over very simple to implement, basic monitoring that catches the typical issues.. because "APM is coming" and "we hired an SRE" and and and..

People want to chase the bright new thing that Gartner is pushing (its always these guys).

> My impression is this is yet another of those Gartner-created hypes...

Bang on!

Its part of same buzzword bullshittery of Kubernetes/ Cloud native / Micro services / observability and so on. Create an out of proportion tech stack that no one would understand and then one needs these APMs to monitor "p99 latencies" because who the fuck knows now on how incoming http request will get processed.

..k8s is buzzword bullshittery?

Look, there's a bunch of bullshit hype, but I think you're being overly reactionary.

Plenty of devs will never touch a profiler (as complex as a cloud based APM simple as jfr), that doesn't mean they don't have use.

Plenty of buzzwords conflated hype and misunderstanding with useful technology

'nosql'-> yeah don't try to do olap on nosql but document databases or kvs are still super useful. It's basically a hashmap with some index what's not to like (also using 'nosql' as a term to differentiate between relational databases and ...non relational is already a misnomer)

'big data'-> stupid hype at the beginning and yeah you can have a functioning product without it but try leading a business without any data to make a decision

'cloud'-> colocated data centers and timesharing have been useful things for decades but man do I love having an API to formalize it rather than calling up the onsite data tech and having a phonecall to do anything (that or a multi day email chain)

'microservices'-> yeah it's intractible if you atomize your app, but splitting application boundaries and defining contracts makes it a lot easier to horizontally scale, and it's cheaper in the end of you do it right. Also functions as a service are great for glue code and infra IMHO.

The longer I get in my career the more I see that wading through the bullshit is needed to know wtf I'm talking about. It's similar to science communications, where the actual researchers will figure out something cool and useful but situational and contextual, and then the zeitgeist completely distorts the findings beyond any measure of recognition

To make it clear, monitoring p99 latency is very often a good thing.

You just don't need a fully featured APM package to do that. People have been doing that for decades, for example by running a script on their server logs.

> "for example by running a script on their server logs" Exactly! There's lots of stuff you can do before you write a $1M check to one of these APM SaaS.

Timestamp each hop. Do your percentiles on a trailing X day window each morning. Alert on worrying moves in the p99/95/whatever.

Make sure you don't run out of disk/memory/database by like.. alerting when they get to 80/90/whatever %. Amazing footgun to not monitor this simple thing.

Have a process that serves as a reasonable proxy for "busy-ness" / bottlenecks. ie - gateways, Monitor the % CPU over X minute windows. If the CPU stays saturated for N minutes, something is bottlenecking.

Here's another secret - you want to also monitor for the opposite. Have processes that are "workers" of some sort that should always have load during business hours as there is always data flowing through? Monitor that their CPU doesn't fall BELOW some threshold (it means something fell over and they aren't doing work).

None of this has to alert within milliseconds. Simply having it, and it alerting within a few minutes is better than not having it. Some of these things like latency trends can be run nightly, whatever.

That's an interns work for a week or a seniors afternoon.

This is easy on s5s (servers). I'm sure you can do this in k8s or whatever too.

Java has an excellent one open source APM - https://glowroot.org/.

Python and other tools - not so much.

I wouldn't say Gartner created hype. Glowroot has been invaluable to us. Of course we run a massive footprint of 0.5m - 0.75m JVMs.

> Of course we run a massive footprint of 0.5m - 0.75m JVMs

Yep, those APM packages are for places like yours. Not for most of the uncountable other places they recommend it.

> fluentbit daemon is hung and no logs came through

Are you talking to me?

There might be pricing pressure, and a lot of people entering the space, but it is more than a nice to have. If you do not have some sort of SIEM solution, you will pay more for or not be insurable for cybersecurity/ransomware insurance.

Implementing observability/SIEM yourself has also become a lot more complicated because the domain of things that have to be observed is increasing. There are many organizations now that have things that need to be monitored in several clouds, infrastructures like k8s, and stuff on prem. So you need a vendor that has the tools to collect this data easily and aggregate it.

The startup world and startup whales might be dying now. But there is a giant iceberg of companies that are becoming more hybrid and their systems that used to do things like funnel syslogs and firewall flow logs to a collector are no longer adequate.

SIEM like in SolarWinds?

Yeah, I imagine insure companies still require this.

Grumpy old developer devils advocate voice: Why does a lot of the modern tech stack seem designed to make you have to pay for a bunch of other SaaS?

Like containers/serverless was supposed to mean "cattle not pets" and it autoscales and autoheals and blah blah blah. And then we break everything into micro services and we need to pay for some SaaS to orchestrate it, and another SaaS to observe/monitor it, and another SaaS to...

These questions are heresies. Just had meeting where discussion was whether system can handle additional 8tps (not 8k tps!) load or they need to stagger it.

> what do people think of the "observability" space from here?

Splunk and other observability companies (Datadog has been hiring like crazy on the security side) have been pivoting into becoming a SIEM 2.0 for a couple years now.

Infra budgets now include Security spend, so an entire generation of Infra companies in segments such as Observability/Application Management (Datadog), Data Platform (Databricks, Snowflake), API Management (Imperva, Wallarm, Postman), and DevTooling (Hashicorp) have begun pivoting into the Security segment.

The same thing happened with networking in the 2010s.

Right I figured Splunks defensive play was to market themselves as infosec which every CTO needs to buy in order to tick a box.

Yep.

It's a 2 birds-1 stone strategy because Splunk can be used by your Platform team as an APM and by your Infosec team as a SIEM. This way you can file 1 PO and use a smaller portion of your budget.

This is was a major reason why Cisco acquired Splunk as their SIEM offering is shit, and there is a massive overlap in customer base.

I work at an observability service and have a personal stake in the outcome so you can disregard my opinion I guess. But it seems to me that observability is there to enable you to get along with fewer people. The more stats and logs and profiles you have the fewer human operators you need.

The savings only really works at scale though right. I've worked in plenty of engineering departments with 200 engineers and essentially 5 operators. You need bare minimum 3 of them for holidays/timezones and key man risk reasons.

So you can maybe help a 200 person org save 0.5-1% of their lower paid staff? Is your bill + engineering integration costs going to be less than $200-400k/year?

Depends on how much the non-operators are using the data. In SaaS companies it is common for the engineering teams to own the performance of their sub-systems, which entails the need for instrumentation, monitoring, dashboards, alerting and investigative tooling. The trend has been to combine things like logs, distributed tracing, APM, error collection and general metrics collection into the same toolset so that you can more easily cross-coordinate between them when tracking down what is happening in a large system. microservices architecture and "devops" make this a much bigger thing requiring every team to take on some aspect of operational capability.

I've worked in 100-person groups where everyone spent all of their time trying to figure out what the hell the system was doing because it did not have tracing or logging.

microservices I take it ;-) ?

Yes it was one of those extremely bad architectures where every little thing that could have been a library function was an RPC instead and all the services were running in containers with .1 CPU quota and 64MB of memory.

Someone needs to explain to me the purpose of containerizing architecture down to such small subcomponents that they could have run on my Windows 98 machine 25 years ago.

At some point you're wasting a lot of time/bandwidth on RPC, and/or memory on all the data copies across process boundaries. So its bigger and slower, harder to orchestrate, harder to monitor, and.. I'm sure there's some positives somewhere.

I don't get it either, and the people who designed and built that system were idiots, so I didn't get a good impression of such fine-grained services. I am comfortable with services that are much larger, such as "store this document" or "classify an email as spam or not".

My personal opinion is that using a SaaS observability service just makes life 10x more easier.

That said, the costs are too high. And in a rising interest rate, declining revenues environment, a lot of services that were previously instrumented are getting de-instrumented or deprecated.

As an example, all the monolith->microservices drama that happened in the past few years is now becoming a questionable engineering direction. Fewer services automatically leads to lesser cloud usage and lesser observability usage.

> what do people think of the "observability" space from here?

I don't think it's cut entirely, but when you're looking at your cloud spend, it's got a target painted on it. It doesn't produce revenue on its own, so people will ask a lot of "do we really need X" or "can we lower retention."

100%. Obs tools are still needed for debugging, incident recovery and just being able to make better products, but there is a lot of optimization going on.

The market still does seem big. Datadog is needlessly very expensive and the experience is just ok.

The experience (other than the sales process) is great.

It's a classic "dev tooling" play that struggles in downturns.

- Company wants to save money, tells team they're cancelling service that costs them $20k/month to observe their production env.

- Dev team starts building small parts they need anyways, chews up significant engineering resources.

- Company now pays more for highly custom solution no one can compete with.

If the company ends up being successful, they'll probably point to all their custom stuff as why. If they end up failing they'll talk about all the boondoggles they went on. This replays so much.

Argument to be made that it may be a better outcome anyway. Custom solution no one can compete with. Engineers usually have some wiggle room to build things they find interesting & useful in their 10% time.

You build the parts that are most urgent as they come up..

Where is this mythical place with 10% time to muck around?

Or is that the 10% of after 9-5 free time you're talking about?

> what do people think of the "observability" space from here?

They're all effectively surveillance systems with extra steps. The Network itself is always listening.

I am still amazed how quickly Cisco destroyed Flip from 2009 to 2011 - basically 590Million down the drain: https://en.wikipedia.org/wiki/Flip_Video

Sure you could argue that by 2011 Flip was on the way out as phones were taking over, but there was plenty of room to pivot and plenty of good will.

Had a relative whose company was to be bought by Cisco but thankfully Flip debacle destroyed that deal.

Flip basically shows up at the same time the iPhone got video. They were dead already and only Cisco didn’t realize it.

Oh Flip was dead in the long run no question about it. It was not only iPhone but Android phones that finally got good enough video quality.

It is just that Cisco cut their losses extremely early. There were still 2-5 years left where Flip could have been profitable.

Without insider information, which I don't have, who knows.

But I will say this:

1. cisco is not a consumer company. they tried with linksys, bailed. tried with scientific atlanta, bailed. no surprise that they bailed on PD.

2. having seen a lot of ASIC acquisitions in my lifetime, which is basically what Pure Digital was - apple knockoff packaging and branding plus an encoder - in my experience tons of startup teams get their first ASIC done, but then totally flub their second (and some of this is the layoff dynamics of how ASIC startups operate on the HW side - unless you immediately start the next chip before finding fit, you have dead weight that you cut). Also, once you show how it's done and have fit, the professionals show up, like Qualcomm, which makes that even harder.

It wouldn't surprise at all if (2) was the culprit.

Despite the iPhone 3gs and HTC Magic offering video recording in mid-2009, GoPro's market cap peaked in October 2014.

It's always 7%.

"T-Mobile is laying off 7% of staff"

"Robinhood Lays Off About 7% of Its Full-Time Employees"

"Sonos lays off 7%"

"Vox Media to lay off 7% of workforce"

"Payments firm PayPal to lay off 7% of its workforce"

"Roomba maker iRobot to lay off about 7% of its workforce"

"GitLab to reduce workforce by 7%"

"Informatica to lay off 7% of its workforce"

It seems like these CEOs are just copying from the same playbook. I wish there was a CEO who was brave enough to ignore the herd mentality, take a look at what the actual needs of the company, and make a bold plan like "fuck it we're doing 8 and a half percent".

“I’ll be there in 7 minutes” is my standard answer when I know it’s going to be more than 5, probably less than 10.

Talk to enough MBAs and you realize most of them are following the same playbook. They all say and do the same things and no one cares as long as things aren’t disastrous. But it is the safer option and probably results in less innovation.

So the 7% cuts will continue until someone boldly breaks the mold and standardizes 8% and brands it “the rule of 8s” and preaches it at overpriced executive retreats.

"Better to fail conventionally than succeed unconventionally" - IMO, this often leads for people making decisions that are defensible by comparison to "industry norms".

Big "No one ever got fired for buying IBM" energy

My guess is that it’s a safe number where you can lay off at that level and guarantee minimal disruptions in operations.

7% equates to laying off one person per group of 14 people. It would be hard to argue that most companies couldn’t get by with 13 people doing the work of 14.

I think you would feel the loss of one person on a team of 10 a lot more.

> My guess is that it’s a safe number where you can lay off at that level and guarantee minimal disruptions in operations.

Only if you fire the people at random. Any bias will change that number.

Bias toward low performers and/or more recent hires would change that number more favorably for the company.

Us peon employees don't like to hear it, but it's true.

Generally, 7% includes most of the cost center divisions (recruiting, HR, BDRs/Demand Gen, Marketing) plus a smattering of low performers in Eng+PM. Speaking from experience in M&A.

Bungie just did 8%

Goldilocks percentage.

CFO's will follow suit and Wallstreet will begin to expect tech to make cuts ...

It's getting noticed that Twitter/X reduced it's staff by 80%, and also reduced it's cloud spend by 60% ... and there hasn't been any material change to their business.

Some might argue that Twitter/X has been able to innovate faster as a result.

Now I'm not suggesting that FANG is going to layoff 80% overnight. But I wouldn't be surprised if CFOs gradually over time continue reduce their headcount, or keep headcount flat over the years to come.

https://www.cnn.com/2023/04/12/tech/elon-musk-bbc-interview-...

https://twitter.com/XEng/status/1717754398410240018

I don't think ANY CEO is looking at Musk's handling of Twitter/X and thinking the layoffs had "no material change to their business"

https://finance.yahoo.com/news/elon-musk-x-worth-less-221637...

and there hasn't been any material change to their business.

The business is worth less half of what Musk paid for it. And that's his estimate of the value.

Twitter's valuation dropped more than 32%+ below the price Elon offered to acquire it for (prior to purchase).

That's the only reason why Twitter forced him to honor his offer.

So it's a mischaracterization to state the value of Twitter halved while on Elon's watch.

https://www.cnbc.com/2022/07/11/twitter-shares-sink-after-el...

> and there hasn't been any material change to their business.

Their revenue has halved. That's pretty material.

Not if the revenue that disappeared was unwanted revenue. <Roll Safe Meme>

I would pay money to never see X used as an example of how to run your business ever again

I agree with you that people are noticing that Twitter going back to the size it was in 2015 has not really impacted the product, but it has definitely impacted the business. Musk's handling of twitter has been awesome for correcting his public image.

I think everyone who is honest accepts that Twitter was massively bloated. By the time Musk made his offer, it was 8x the size it was in 2011 and the product really hadn't changed much in a way that required that kind of increase. But Twitter is an awful experiment because of Musk's clownish handling of just about everything.

This is bad news because it happened before the acquisition. When Cisco buys them, there are going to be more rounds of layoffs.

Well, never the less, these deal are all done with a set of vested interests that in many occasion don’t necessarily align with neither the customer’s (why companies should even exist) or that of their employees ( the motor of companies success) but more in the realm of financial spreadsheets, where, due diligence sways in the direction of those stakeholders in the deal… And I purposely look at it this way as no to take any given side without the details.

Non the less, IMO, the only winners here are the stockholders, Splunk as a business, and many others with the same model of “schema selling” are in a high risks stake in the new erra of AI/LLMs.

If you consider the accelerating world of AI we are living in, and the emergence/trend towards Domain Specific Large Language Models (DS-LLMs) and advancement like MEMgpt, they represent a transformative approach to data analytics. Instead of using a schema-specific model, as seen in tools like Splunk which extract and transform data into a predefined schema, DS-LLMs offer a flexible, continuously trained approach. They not only analyze data but also learn from it in real-time. The “actors”, or bots, leveraging tech like MEMgpt that not only collect but also learn from the vast streams of data are far more capable than those schema models. As these models self-train and trade knowledge, they are poised to provide insights more organically aligned with the data’s inherent structure, rather than a pre-defined schema. This means businesses could potentially gain deeper, more intuitive insights without the confines of structured data models. With the rapid pace of innovation in the AI sector, it’s worth questioning whether traditional, schema-based solutions will be able to keep up with the dynamic learning capabilities of DS-LLMs. I still wonder who got the better deal here.

Wishing the best to all the Splunk employees moving forward.

Do you need hackers for hire? Do you need to keep an eye on your spouse by gaining access to their emails? As a parent do you want to know what your kids do on a daily basis on social networks ( This includes facebook, twitter , instagram, whatsapp, WeChat and others to make sure they’re not getting into trouble? Whatever it is, Ranging from Bank Jobs, Flipping cash, Criminal records, DMV, Taxes, Name it, We can get the job done. They are a group of professional hackers with 10 Years+ experience. Contact at ( remotespywise @gmail com) … Send an email and Its done. Its that easy, try us out today.

Company gets acquired. Company lays off employee. Typical.

Yes, any acquisition is a signal to look for a new position. Your cheese has moved.

And for some reason we as employees continue to let companies abuse us, despite having the option to fight back.

I am so bitter and jaded when it comes to this entire system.

Fight back by starting your own company then

And the cycle continues. Maybe they are talking of unionising

I am talking about unionizing. Workers deserve better than what they've got and the deserve the protection against a company a union brings.

The company is not acquired yet.

>Cisco buying your company should be considered a worst case scenario for most employees. The important thing to remember is not to trust a word they say.

Absolutely true.

The past year and a half has revealed that a lot of programmers have the same "sense of entitlement" as laid-off Ohio steelworkers or West Virginia Coal miners.

I'm not saying you shouldn't complain about your employer. But the notion that companies are ethically obligated to hold onto workers who aren't necessary, I thought we were way passed that.

That entitlement was always there, it was just papered over by a boom driven by ZIRP. Now that market conditions are less favorable, perhaps that entitlement might actually be channeled towards unionization activity.

It would be cool if there were laws or something requiring the acquiring company to keep all employees on the payroll at their pay rate when the acquisition occurred, for a minimum of a 18-24 months. Sadly the majority of our politicians are glorified fluffers for big corporations, so nothing like this would ever happen.

With that plan you may as well just make M&A illegal. The net result would be 0 mergers other than giant companies acquiring small startups with minimal payroll.

Unfortunately, it's how M&A works since forever, even if the official statement is "The overall market has retracted". I guess most of the cuts are in the "non-productive" departments like sales, marketing, hr and stuff like this.

It's not that they're 'nonproductive', per se; that's kinda pejorative. And it's not the 'market'. It's that you just don't need 2 HR, accounting, marketing, finance or other back-office departments, post-merger, so most of those folks become redundant. The acquiring company is going to convert to their systems asap anyway; economies of scale mean you don't need as many bodies to cover the acquisition. You usually try and keep a good percentage of sales guys if for no other reason than there are deals in flight you don't want to loose, but eventually those guys get whittled down too, usually right after the change from "small company fast and loose comp plans" to "corporate comp plan same as everyone". And, of course, the execs have cut their own deal. It's when they start cutting core engineering and support right away that you go "wassup with that???".

In the cold calculus of M&As, 7% sounds like an entirely normal number, IME.

I remember vividly when one of my companies got bought by IBM, and a couple of the younger back-office folks came by to ask me "what do you think?" because I'd been through M&As before. "I think you need to look for another job" was the best I could do, and it was a little disheartening.

Does anyone use New Relic here? I've seen a couple of jobs posted from them on LinkedIn recently but not sure of the companies current status in the market. Where does it fit in against splunk and datadog etc?

New Relic was amazing when I used it a few years ago. One of the few products I’ve been seriously impressed by. It is expensive but seems worth it. They had very good support for async tracing, much better than splunk and Open Telemetry at the time.

Their graph and front end tools were just seamless to use. It was a very well done product.

This seems to be the impression I've gotten from the little research I've done but other names are ringing out more in the general chats I've seen, I'm just not clear why.

They're also in the middle of being acquired by Fransisco Partners and TPG. I'm not sure what to expect there... cuts or restructuring would leave me pretty exposed if I moved there in the recent future but it seems like a good place with a bright future... Decisions decisions!

They delisted and got acquired by Francisco Partners and TPG recently. They aren't doing that hot because Datadog and (surprisingly) Splunk has a stronger vision. They have a pretty nice rolodex though, so it's a good PE move. Bad as an employee.

Oracle should have bought them. They are the ultimate rent seekers.

M&A is noncompetitive and should either be prohibited or heavily taxed.

Create a subsidiary? Fine... Buy up all the tadpoles so you don't have to deal with frogs? Noncompetitive...

I don't totally disagree but the consequence would be 80% of VC investment vanishing overnight and a whole class of founders just not bothering.

The chance of a successful IPO is faaaaar less than "bigcorp wants you to go away, so here's some $$"

Where would the VCs put their money instead? NFTs?

VCs are given money by larger investors to play around with finding the next Oracle/Facebook/Uber/etc. If VCs have nowhere to invest or VCs mess up a fund, LPs will ask for their money back and invest elsewhere. This is what happened to Softbank, though they ended up having the last laugh, thus cementing the fact that Mayaoshi Son is our industry's Sheogorath.

https://archive.ph/gYGDv

Anyone has any info on any specific teams that got cut?

You can check out on Blind. This round seems to be pretty senior-level heavy. Layoffs mostly hit the senior/principal engineers.

only 7%?

Reply on news.ycombinator.com