Hacker News new | past | comments | ask | show | jobs | submit login
PagerDuty S-1 (sec.gov)
333 points by johns 7 days ago | hide | past | web | favorite | 187 comments





> If we fail to offer high-quality support, our business and reputation could suffer.

I don't know how common this is in S-1 filings but it stuck out to me. Support is often viewed as a cost center, I love that PagerDuty makes it such a high priority.

Edit: After some research, it turns out this is common. Docusign, Dropbox, Okta, Zuora, and MongoDB all have this statement exactly or similar.


I can vouch that Support & Customer Success has always been a first class participant in the company. One of the founders ran the team when it was small, and Ryan has done a great job growing it out -- to the point where their #1 problem is other PD teams poaching their people into other roles in Product, Internal Tools, Engineering and even Sales :)

(Disclaimer: I worked at PagerDuty for 7+ happy years)


> Support is often viewed as a cost center

Support is a cost-center for B2C companies, but one of the most important parts for B2B.


Part of what Businesses pay for is support. Heck think of all the free databases/open source stuff that is funded not by the code but buy offering support/integration.

EDIT: Funny though, as a consumer I'd pay a little more, say 20% more for my comcast bill if they had what I'd describe is "nice and helpful support", but if that was the case, my bill would likely be accurate, and we wouldn't have all these BS fees and charges showing up out of nowhere, which coincidentally can make my bill 20% higher then advertised.


You can get Comcast Business service. That's basically what it is. They have very good support but costs a bit more.

It’s still usually viewed as a cost center by management in either case, and employees in support roles are notoriously underpaid, overworked and at risk of inexplicable layoffs.

I would love to see more companies treat support technicians and customer support agents like prized human capital, give them good job security, don’t try to outsource or downsize them, and give them top class compensation to make them very proud to do the job.


Which is why managers in B2B support organization often start moving toward profserv ... which then creates a downward spiral in the value the core product provides.

Google would disagree.

Support is a cost-center for most B2C companies.

I would consider it a cost center like a developer fixing a released bug. It's necessary maintenance and when done right results in more profits

This is another reason why everyone at a company should be a csr for a number of hours per month. From CEO to every developer.


Support as a cost center...it's an interesting thing. My first computer related job while in college was support for a variety of Borland products. There were a few of us interns that ended up doing the work of senior support engineers - writing white papers, running training, etc. Interns were the "cost cutting measure" back then.

We were paid less, but knew the product and some evolved into well paid and challenging jobs with the company, some of us went elsewhere.

Local labor was used instead of outsourcing overseas or bringing in lower-paid workers.

At the time interns were paid pretty darn well.


Im of the opinion that old support people make the best QA people and know quite a few you have made the transition very successfully.

True... but to be fair, prob most of the companies they serve, their support depends on PageDuty to be working correctly (othewise they would not get the alerts/notifications). So I would expect PageDuty to mark this a super high priority.

More and more companies are also involving PagerDuty in their support! Incidents and multichannel escalation are great complements to standard support workflows for high impact cases.

Professional marketer perspective:

Sales and Marketing is around 60% of revenue at 47 million a year. Advertising is around 5 million a year, so the rest of the spend looks like sales costs, content production, and maybe event marketing.

The media spend seems to be mostly adwords: "The effectiveness of our online advertising has varied over time and may vary in the future due to competition for key search terms, changes in search engine use, and changes in the search algorithms used by major search engines."

Also, they only have 5000 FB likes - if they were spending $100k+ a month on FB these likes would be higher as a byproduct of the paid spend. I do a ton of B2B/SMB FB advertising, my guess is that they could move some spend to FB from AdWords and see better ROI.


Jesus, do people really like company FB pages (especially B2B companies) at all these days?

What's FB?

Facebook

If I see something of interest I'll like it just to keep an eye on it. If after a couple+ weeks it doesn't pan out, I unlike.

For every $ spent in fb newsfeed ad spend, you'll get some folks liking your page.

Real, or fake people? Do those likes quantifiably translate in to $$s?

Real, and not really. It's a side effect of generating sales.

Could they really see a better ROI moving from Adwords to FB for a B2B market? I'm curious, because I run a B2B business and I never would have thought FB ads would work. I always assumed FB was for B2C. I currently use Capterra and Adwords.

we tried it and it did not work for us, sure we got some leads/registrations from facebook, but most of them registered for whatever reason, im pretty sure they did not understand what they are signing up, and those that where ok (maybee 10%) did not stay after trial. we even tried traditional media/portals banners with a bigger budget, nothing, and for both of those we hired outside agency to run it. The only thing working for us is adwords/seo and recomendations. it would be great if we could diversify but for our b2b it does not work like that.

i can imagine fb being great for b2c especially for something with emotional connection. but for b2b i'm sure that in most cases it does not work, at leas not in creating leads/registrations.


I can guarantee you it can work for B2B at scale. I've personally grown multiple B2B accounts from $0 to an ROI positive six figures monthly. From seed stage start-ups to post IPO companies.

That's not to say you did anything wrong - it's possible your specific set-up wouldn't work. But many agencies do not know how to do it.


how is that measured ?. initially we thought, how stupid we where in not paying for facebook earlier because of increase in signups. but on closer look the registrations all where from private individuals (remember we are b2b) and they would never fill out the data about the company or even log again after initial signup. even those that where registered as companies never used it.

i think facebook user has a different expectation then google. i know some companies first hand that had great expirience with b2c, i never meet some that did great in b2b.

maybe if our focus was more on creating a community around a common thema, and then do brand building or so, it could work as a secondary generator of leads, but for pure,i pay for an ad and the goal is to get an conversation it was not that great.

so if you could write more about what does work it would be great because only relaying on adwords and always increasing cpc is not that appealing.


Measured by revenue generated per $ spent. What works depends on your specific situation.

I'd recommend both. For example, social media is a good place to do retargeting because people spend so much time on it. Usually after you've warmed up potential customers either on FB channels or outside. Also, Facebook owns a lot of mobile engagement, which is pretty huge.

50% of all americans, and thus business owners, use FB or IG every day of their lives. I've personally managed at least 10 mil in profitable B2B FB spend.

The Facebook advertising spend is most likely remarketing, not post/page promotion.

Could be. We have used Facebook ads to provide branding cover at events. You set a super tight geofence for the campaign and saturate everyone at the conference. The modern equivalent of putting ads in the conference bathroom.

I'm sure they do remarketing. I'm saying that their pure acquisition spend on FB doesn't look high.

Edit: They have no live ads, so they are not running FB retargeting even:

https://www.facebook.com/pg/PagerDuty/ads/?ref=page_internal


Am I reading it wrong? They don't fill the data in section MANAGEMENT’S DISCUSSION AND ANALYSIS OF FINANCIAL CONDITION AND RESULTS OF OPERATIONS:

....The majority of our revenue is derived from midmarket and enterprise customers. As of January 31, 2019, we had more than customers globally, with [missing #?] customers having ARR in excess of $100,000, and [missing #?] customers having ARR in excess of $1,000,000. Our 10 largest customers represented approximately [missing #?] % of our revenue for the fiscal year ended January 31, 2019, and no single customer represented more than [missing #?] % of our revenue in the same period, highlighting the breadth of our customer base. We serve a vital role in our customers’ digital operations and grow with them as their needs expand. As such, we have developed a loyal customer base, with total ARR churn representing less than [missing #?] % of beginning ARR for the fiscal year ended January 31, 2019. Our ARR churn rate represents lost revenue from customers that contributed no revenue in the measurement period but did contribute revenue in the equivalent prior year period. We generally bill monthly subscriptions monthly and subscriptions with terms of greater than one year annually in advance.


There are many fields that can be left blank when filing an S1. They get filled in as time moves on, but need to be complete before the IPO takes place.

What's the point of filing then? It sounds like absolutely critical data.

I guess the people who they want to know will have all the information already. Nowdays S-1 filings can be confidential until three weeks before the IPO, many rules of the market do not make it a even playing ground for retail investors.

thank you.

I must admit I didn’t read it, but that seems like that would be pretty embarrassing. Maybe it’s not required? Idk, seems like a huge oversight for any company, let alone one that seems to have their shit together.

Congrats to the whole PagerDuty team!

We were in the same YC batch (S'10), and the founders were all kind, smart people. They were also determined and hard-working: I believe they'd been rejected by YC three times before getting in.


To save anyone else from staring at the page for a minute (wondering if it's linked just because it looks like 1992 called, not knowing what the "sec" USA government department does or what an S-1 form is): PagerDuty seems to be going to the stock market.

Thanks, I was quite confused.

PagerDuty is super expensive. With Slack or another messaging app, you just need a system that will support a schedule and route to engineers appropriately as well as post to the slack channel. A lot of the bells and whistles they have aren't useful. Engineers never need to actually use the Pager Duty website and only use the app for ack'ing the page or just do that in Slack.

This feels like it should be more of a commodity. E.g. a service like OpsGenie can undercut them.

$100 a month per user on their enterprise plan is insane for what they offer: https://www.pagerduty.com/pricing/


What industry are you in where you care about getting paged after hours, but you're happy tying your paging to Slack's availability, and you have the engineering time to devote to recreating and maintaining a paging system?

I'm sure there are industries where that makes sense, but I bet you can imagine there are industries and companies where that doesn't add up.


You still need something like PagerDuty, OpsGenie, etc. for the on call rotation and routing. All I am saying is all the other 1000 features they offer and charge for aren’t needed when you don’t really use their web site at all.

Also, my company switched from PagerDuty to OpsGenie and after a week everyone was used to the new system, so that strikes me as more of a commodity.

We reimplemented paging system with Slack plus Twilio fallback, it made sense for us, but yes I think that's rare.

Yes, Slack only was completely unacceptable and we already had Twilio account used for other things.


I haven't created my own paging system but I've gotten quite far with Zabbix and Twilio. Zabbix's feature set includes the ability to define teams and schedules and execute a custom command when downtime is detected.

That being said, I was happy to hear that my new work had a PagerDuty account because the fewer things I have to host and maintain the better.


SRE at a large software company here - PagerDuty is absolutely critical to our operations. I will absolutely not trust Slack's availability nor is Slack the right platform to manage this and to expect immediate responses, even from oncalls.

When you've lots of users bells and whistles like schedule layers and setting overrides are very useful.

You don't need the $99/month plan, you can do fine on $39/month.


I didn’t say replace the “operator” part of PagerDuty — you still need something to manage on call rotations and escalate to right engineers. But that shouldn’t cost $39 a month per engineer and all the other features won’t be used because people spend most time in their messaging app triaging issues.

I see you've never had an issue that needed someone to immediately and reliably answer. Even I've had to rely on PagerDuty's ability to both persist in calling and fallback to the next person (all with a history to audit) when a super serious issue occurs. The key though, is for PagerDuty to only be invoked very rarely, so that it isn't becoming abusive.

These routes are configurable using Prometheus.

Yes, and Dropbox is still trivial for Linux users to build. But somehow, they're doing okay post IPO. If you think slack is a viable alternative and are complaining about sticker price on the enterprise plan, perhaps you are not in their target market.

Why are you citing the enterprise plan if your argument is that you don’t need most of what they offer?

It’s $30/month for lowest tier with more than 6 users. There’s no way it’s ROI positive for you to build and maintain this yourself until you have [many] hundreds of employees using PD.

The fact that it’s super simple to use while being highly configurable and reliable is a pro, not a con.


There are cheaper services that do same thing. If you aren’t living in PagerDuty to do actual incident management, but only use them as an operator, it is hard to justify 360 a year. That’s way more than even something like gsuite.

Have you actually tried this? I did, and Slack's availability is abysmal. I implemented fallback just in case and got surprised how often it triggers.

$100/month if you want the things that offers (and I imagine many enterprises don't need many of those things). Also consider that's the list price. There is a reason they have sales people. I'd be surprise if most people are paying list price.

Exactly why we are building https://amixr.io

Schedules, lifecycle, routing in Slack.


Still $29 a month though…

Charged only for on-call persons. So should be less comparing to other tools.

Example: if you have 50 engineers and only 2 of them are on-call per sliding week, you'll be charged for 2.


Ohhhh, that is much better. Totally agree that is fair given there is a cost to actually page someone.

Nice work!


Thanks!

Interesting that they reference their Glassdoor rating:

The strength of our culture is key to attracting and retaining the best talent, as demonstrated by our high employee retention rates, and, as of January 31, 2019, a Glassdoor rating of 4.5 out of 5 and 100% approval rating of our chief executive officer.


Uh, that is a big negative sign for me, because they have an an "employer active" account. In other words, they pay for the privilege to get reviews deleted.

> 100% approval rating of our chief executive officer.

So I am 99.99% sure that HR was tasked with removing bad reviews ;-)


A previous employer of mine had a Glassdoor "problem" with bad reviews. Now when people leave they are bribed with a week's pay to promise not to post anything bad, and their Glassdoor reviews now are packed with glowing praise that look quite fake and contrived.

Show me a rating system and I'll show you ways for people to game it.


Is there any evidence that glassdoor lets employers do this? I think glassdoor actually specifically says that they won't, unless it violates a law somehow. They even specifically state that they "don’t take sides in factual or contractual disputes between employers and reviewers."

https://help.glassdoor.com/article/I-m-an-employer-What-can-...



Anecdote: A former company I worked for had a negative Glassdoor Review. It "disappeared". Dunno how, but it's not there anymore.

I suspect at some point they could be removed. But again, anecdote.


Generally employers figure out who left the bad review and threaten legal action to the person who made the review. It isn't Glassdoor that removes the review but the person who wrote it.

Also anecdotal: our employer "highly encouraged" us to write glowing reviews of the company as a marketing effort to make the company look better to potential hires.

Potentially leaving them open to the pettiest securities fraud suit ever.

That's so unfortunate. I remember back when Glassdoor prided themselves on not doing that.

Oh look a penny..

How does one know when an account is "employer active"?

I was checking their careers page (because.. you never know!:) and they have the 'courage' to link Glassdoor on their careers page[1]. Perhaps other companies also link Glassdoor to their careers page, but this is the FIRST time that I see this.

[1]:https://www.pagerduty.com/careers/


Pretty funny considering that Glassdoor will delete negative reviews.

I've read some pretty negative reviews when searching info for a company. As long as it 'looks & feels real' they it let them. Someone already posted the glassdoor takedown rules and they seem reasonable.

I'm a pagerduty user and I've met a bunch of their people. Their service is top notch and their people are really cool, knowledge folks. https://response.pagerduty.com/ is a _MUST_ read for anyone remotely near operations or support.

2017: $79.6M revenue, ($38M)in losses.

9 months of 2018: $84M revenue, ($34.5M) in losses.

Costs seem generally under control except the significant growth (+66%) in G&A expenses.

Why the hurry to IPO?


My mostly uninformed opinion is that smart money smells blood around the corner and wants to cash out before it gets bad. Mass retail store closures, layoffs, IPOs, ...

Even if true, this will largely not affect a business like this. They make small money on each company and scale through the shear number of total companies that pay a few thousand dollars a month to use this service. People are not going to stop going on call because their revenue dropped 20%...

The market is still hot. You don't want to go IPO at say $50 and no trades. When the market crash everything will crash along with it. Some investors are probably very eager to cash out now than later.

Liquidity for one.

Conditions for an IPO are rarely perfect - you need to have the right internal situation, the right market conditions etc. etc. - so if the company is thinking of doing an IPO, and they can ... well then they're going to lean towards doing it.

VC's can cash out if they want and pass the buck.

That said, if they only have some major 'one time cost' dragging down earnings, and earnings will be spectacularly better next year ... it might be worth the wait. But even then - if the IPO suffers a little, but then rockstar numbers come out later, the stock will go up. So VC's that want to hold on have the option to do that.


Good market now, bad market likely in a year or two.

Why do you think bad market likely in a year or two?

We're way overdue for a recession, there are already signs of slowing, and historically after massive tax cuts for the rich, a recession follows a couple years later after the piper has to be paid. Signs all point to a recession in one to two years.

> Signs all point to a recession in one to two years.

People have been saying this for the past 7 years.


They're bound to be right some time though, as always.

Maybe, but people who actually know what they're talking about have only been saying it recently.

> citation needed

A year ago the markets zeitgeist was "synchronised global growth". Then it was "narrowing term spread on bond yields". Now it's "China and European growth slowdown" and "US Q1 earnings softness".

Getting off the news headlines, market prices for the last couple of years were broadly consistent with Fed interest rate projections -- slow but consistent rate raises last year and in 2017. That is no longer the case. Now the Fed has backed off the pace of raises and is (IIRC) projecting zero or one for the rest of the year, and markets are projecting rate cuts for the first time in a long time. If people think the Fed will need to cut rates, it's because they think inflation will be down, likely due to lower wage pressures from a softer labour market etc.

Other things: some market indicators have "gotten better". P/E multiples and similar metrics were historically high and have trended down (both due to price drops and earnings increases.) Blame interest rates rising last year IMO, why wouldn't ratios be high when rates are low? They're still high, but not worryingly so.

TBH I don't give much thought to the Chicken Littles. Maybe we'll have a technical recession this year or next, maybe due to trade, maybe due to China or Germany, but for the moment I think people in the markets broadly agree that things are long-run pretty healthy. Famous last words, I guess.


What's the trigger?

Pension, student and auto debt, I'd guess.

It's worth noting that since the collateralized housing debt scandal of the Global Financial Crisis, approximately nothing changed in regulation, and credit card and auto debt securitization has taken off in its place. Investors still want to put their money somewhere.

Who is the Lehman Brothers of the new securitized debt? That meltdown is coming.


Unlikely, either because of guarantees (student) or too small (auto loans). It’ll be the corporate bond market getting spooked, and a crunch when anyone running on cheap debt can’t service or refi it without default.

US auto loans have a total of 1.1 trillion outstanding, writing that off wouldn't affect the markets at all or increase the cost of debt?

That collateral is easily repo’d and disposed of. Think Rent A Center. The margins on the vehicles being sold make up for estimated losses.

https://www.pionline.com/article/20181101/ONLINE/181109968/l...

https://www.mercatus.org/bridge/commentary/pension-crisis-yo...

https://www.bloomberg.com/news/articles/2018-10-17/the-stude...

https://www.citylab.com/transportation/2019/02/subprime-car-...

I think multiple sectors of underfunded pension, student debt and auto debt all coalescing at once is the problem, not any one of them individually. You're also negating that folks defaulting, massively hinders their ability to move in the financial system. And yes, you can put cars back into the system the same way you can put houses back into the system, but it doesn't matter if no one wants cars.


> You're also negating that folks defaulting massively hinders their ability to move in the financial system.

Quite the contrary. Lenders will lend to you immediately after bankruptcy because you can't default again for seven years, for example. If you have bad credit, you can still get credit! Just at a higher interest rate. I am familiar with lenders who will get you a mortgage the day after foreclosure for 200-250 basis points over conventional rates (6-6.5% versus 4%, in this case).

Economic drag due to student loan debt that can't be discharged is a separate issue (with those debtors delaying or opting out entirely of home ownership and/or children). Pensions will get dumped onto the Pension Guarantee Corp with a reduced benefit payment to those entitled to such, and any federal debt will get inflated away through the Fed (not great, but what happens when you have your own central bank).

As I said, I think the only "great unwind" or credit crunch in the near future is corporate bonds.


What lenders you are specifically referring to?

Subprime credit cards: https://www.cardrates.com/advice/list-of-subprime-credit-car...

Subprime auto lenders: https://www.cyberleadinc.com/subprime-auto-lenders/

Nonprime mortgage lenders: https://www.nonprimelenders.com/lenders/

Prime = good borrower. Subprime/nonprime = not a good borrower.

Just quick examples, lots more out there. It is incredibly difficult not to qualify for credit entirely. Someone, somewhere will lend to you at an interest rate in accordance with your risk profile. Investors are hungry for returns.


Oh boy, ok. Just making sure that's what you'd say. Good luck out there pops.

Compared to earlier periods in history, us stock market prices have been very high for past few years. this doesn't necessarily mean us stocks are overvalued -- maybe all alternative investments have similarly high prices, or equivalently, low yields. But, lots more room for stock prices to go down in future than to go up.

Ignoring hand waving comments about market prices, re: us economy, arguably we might be about to see an upward trend in the unemployment rate, which is a leading indicator of recession, although perhaps part of the increased unemployment in the last published statistics was caused by the government shutdown.

https://financial-charts.effingapp.com/usa-unemployment


> Why the hurry to IPO?

ITT: "A profitable tech company going public while there is still growth potential? Is this a scam?"

Aside from the liquidity and valid option of getting a funding round from underwriters has this seriously become a foreign concept in this market?


It's expensive, it's a hassle, some notable founders have publicly said they regret going public. I think letting insiders cash out is a great thing, but some disagree. From an outsiders perspective it seems like there's plenty of private liquidity for promising, growing businesses.

Seems reasonable to ask why they'd do it to me. I certainly wouldn't assume it's a "scam", but I don't think going public is the obvious good that (I hear) it once was.


Yes having to answer to the public markets and considering control issues of your company's shareholder composition can be a distraction.

I think you are underestimating what liquidity means. Liquidity means having access to the cheapest credit markets on the planet: commercial paper. Having access to the rest of the corporate bond markets too. Being able to dilute your shares and sell them at will. Being able to attract employees with tax efficient competitive compensation packages, again via share dilution so for free. People aren't just saying "liquidity" so that insiders can cash out of the stale balance sheet line items that they update once every 18 months.

A profitable company at a low valuation and larger addressable market is also a rare opportunity for non-accredited investors, solely because of the attitude of treating the public markets as an anethema and just using it to dump on everyone's mom's retirement account.


Half the growth in G&A expenses is due to their one-time donation to the Tides Foundation. If you back this out their expenses growth here is in line with everything else.

Their seed round was in 2010—early investors are likely pretty interested in an exit

Not to mention early employees

employees have no say on this matter

VictorOps was acquired. I imagine this creates various pressures.

PagerDuty feels like abandonware. It hasn't changed in the last 4-5 years.

How much can it really change though? It’s a fairly simple concept. I always considered the cost is paying for trust and reliability. I know a text message might fail. I want those alert redundancies and I don’t want to have to create that system.

I tried some of the competitors out to reduce cost, but UX was definitely lacking (opsgenie for example). We stuck with PagerDuty.

That being said, as a potential investor, I don’t think it would be hard for these companies to reproduce what PagerDuty does, or anyone. They just happened to be one of the first but companies with large pockets, especially Atlassian/VictorOps can make a lot of headway here.

I was surprised to see they aren’t profitable, but I guess a lot of the spend is marketing/sales and the long cycle to get those enterprise customers. Especially the $100k ARR ones.


The last time I looked at PagerDuty (a while ago) it didn't have any kind of "calling tree" equivalent for validated notification escalations.

Couldn't really see how that could be true, so I'm assuming I didn't understand the product.


We use multiple schedules with Pagerduty, and when an escalation policy is triggered, it can route to multiple additional schedules. So let's say you can escalate to an SME schedule, if a certain incident is not resolved within a certain amount of time.

They have a configuration tab for "business services", where you can set up dependencies between services, but we're not using that, so idk how smart is that.


When I call my insurance or bank, I get a pleasant voice interface.

When PagerDuty calls my carphone (or worse, my motorcycle helmet communicator), I get an inhuman civil defense robot from 1980 that only accepts touch-tone inputs.


Better than old school pagers from the 80's which just had an earth shattering sound.

I remember being on call for Telecom Gold (uk version of Tymnet) got home from London late pager fell down the side of my bed and I didn't notice.

Later that night at 3am it went off like a small nuke and it literally took me 20 mins and a coffee to gather myself to be able to call the Operators and then logon (dial up to a x.25 PAD)


But, it absolutely never fails.

This is not true in my experience. I led the monitoring team at my past job and we spent a non significant effort monitoring PagerDuty due to often delays in ingestion and notification. If you page through their status page[1] they have a problem with delayed ingestion or notification almost monthly. If people actually monitored they'd be surprised how often they miss SLA.

[1] https://status.pagerduty.com/history?page=1


I don't know if I would go that far, but the service is quite reliable. I have experienced some spot issues with message and alert delivery over the years, but overall, reliability is strong with this tool

I wish I could opt in to that for all automated communications. Fighting my way through voice-recognition menus is always a pain in the ass that makes me wish for the telephone menu systems of the 90s.

I like having both options!

When Pagerduty calls me, it's because I've missed the push, email, and SMS notifications. The phone ringing gets my attention but I usually hang it up and open the app.

If I'm driving, secondary oncall can get it.


They're spending $30M+ on R&D, so if you feel that way as a customer, that's not good.

On principle alone I wouldn't give them my money ever again. I'd engineer my own lightweight solution before then or use an underdog competitor. I have a sour taste in my mouth for their meteoric rise to success/IPO versus their pathetic investments (or lack therof) in product improvement. It really sucks to use. Reminds me of Engineyard (also abandonware).

I feel that while OP is correct - PagerDuty hasn't changed much - engineering my own on-call system is the last priority of pretty much any business. It's one of the things you don't want to do yourself.

Additionally, PagerDuty works extremely well and I can't think of any additional features we need right now.


Honestly, if your response to critical ops notifications is to engineer a lightweight solution as a part time pet project, I don't want to work for or use your product.

If you are talking instead about hiring an ops engineer to build your custom lightweight solution, I would ask why you're spending 100k+ all-in cost on an engineer to do something you could pay PD to do for cheaper.


We tried using one of their competitors and strongly disliked it, enough to go back to PagerDuty in under a year.

I'm sure there are other options we could have tried, but it seems to do what we need pretty well.


Well compared to victorops it's 5 years ahead. We recently switched off victorops to PD. The victor-ops scheduling UI is crazy clunky. Also PD seems to page faster, and the mobile app is better.

I went from a company that used PagerDuty to one that used VictorOps. Exactly as you say: the PagerDuty UI is much better than the VictorOps one.

For me it's mostly about changes to their phone/watch apps and schedules management

And check this https://www.pagerduty.com/whats-new/



If you want a really crappy experience, try something like Everbridge. It’s a platform for calling parents to tell them today is a snow day masquerading as an IT alerting platform. You’ll be begging for PagerDuty in weeks.

Agreed, too many good open source solutions now. No one thinks of alerting as paging anymore. You can use email to text now, chat solutions like Slack/Mattermost/etc. You no longer need an SMS gateway and if you do then Prometheus supports Twilio. All PagerDuty has going for it is a sales team trying to get managers to sign up for a service before exploring alternatives.

I swear half the time my Slack notifications don’t go off on my phone. Even with background refresh on. I would never trust that service for critical alerts. As in, you need to wake up and fix this now.

I’ve run my own monitoring solutions. My own Asterisk systems. Even my own Twilio integrations. You are grossly oversimplifying what is needed to create and maintain those systems. For many, PagerDuty is absolutely worth the cost, which is why it’s so popular. People will pay for reliability they can depend on so they can focus on their core business.


I think by default slack has Do Not Disturb enabled from 10PM to 8AM. May want to double check that!

Figured that was it, but it’s happening during my daytime outside of the DND.

Slack notification delivery is definitely not what it used to be.

You mean to say it's slacking off?

The value of Pagerduty is not in its notification, but the workflows around it. Alert in the middle of the night? Don't page unless its critical. Management of on-call schedules, escalate when on-call engineer doesn't answer. We use OpsGenie for this (in addition of sending messages to a chat, but off work that is usually muted or not installed on engineers phone), but PagerDuty has the same functionality.

Out of genuine curiosity, do you have a list of good alternatives or some reading material on "paging" best practices? My team uses PagerDuty but I've never really questioned if there's a better method out there because the system PagerDuty replaced for us was so terrible that PagerDuty seemed like a godsend.

Google's free SRE handbook has some info.

You guys are out here giving very strong dismissals of a very good service, and then referring us to a resource you wont even bother to link? Maybe you can link the actual resources to prove your argument instead of assuming you are correct and dismissing people to figure it out on their own. Usually, the standard for conversations here is held to a higher, non-twitter level, and I for one enjoy conversations held to higher standards.

I think they must be referring to this: https://landing.google.com/sre/sre-book/toc/index.html

I don't see alternative services on this page (https://landing.google.com/sre/sre-book/chapters/monitoring-...) Do you have any alternates you would recommend?

What are they doing with all those R&D dollars?

Timbits and Crown Royal. It’s not going to the product that’s for sure.

FWIW, you are close, but substitute some sort of bourbon or scotch instead of Canadian whiskey :)

EDIT: this was a joke. I worked there so I know where the R&D dollars went: to quality engineering. But it is the damned truth that we drank Bourbon and Scotch, not Canadian whiskey


the fireball was canadian (though maybe not "whiskey") ;)

I've been a Pager Duty user since 2011. It's a good product and fulfills its need. I don't see it as a product requiring 8 figures of R&D, so seriously what are they doing with it all?

Man, if I was them I'd be going for Bearface Whiskey and Pie Commission beef pies. But that's just this humble Torontonian's opinion.

This makes them probably the second ever YC company to go public, the first being Dropbox. Also related: https://news.ycombinator.com/item?id=758653

Super quick napkin math on YC's investment. Many of these assumptions will be wrong.

    YC investment: $120,000 [0]
    YC shares: 1,050,000 [1]
    PD market cap: $1.3 billion [2]
    PD share price: $20.23
    YC investment value: $21.25 million
    YC ROI: 175x
This is a ~4x return on the S10 batch. In that same batch you have Stripe, which looks like it could be a 3000x+ return for YC.

Notes

[0] - This could be less, I don't remember the dates YC started messing with this.

[1] - The 3 founders hold 13,670,067 shares, or 91% of 15M shares. I assume YC has 7% of 15M.

[2] - I assume their Sep 2018 raise valuation as the price. I'm not confident the public will value this the same.


Thanks for this! Love to see quick math.

How do you account for dilution though? I’m not too familiar with how it all works, but in this case does it not matter because the amount of shares stays the same? Only total shares have changed?

And yeah Stripe will be a crazy multiplier that has to be at least 2000x and could be over 3000x like you said. I’m guessing they never sold any shares. Especially seeing Square’s market cap above $30B and PayPal at $120B.


https://news.ycombinator.com/item?id=758653

its cool to see how far they've come


<rant>

I run a service that integrates with pager duty on behalf of our users, and man oh man did they make some poor design choices with their API.

The most glaring one is their integration keys, which is how you trigger an alert. They’re just v4 uuids with the hyphens removed, or in other words, completely random.

This is fine by itself, except they have two versions of their event API, and keys for the v1 API don’t work with v2 and vice versa. And there’s no way to distinguish what version a given key is for, since it’s just a jumble of bytes. We can’t even look up the service integration to find out what version it should be. We can using yet another different API, but even then it’s not clearly distinguished.

So now our users are confused about what version they should use. We only supported v1 for a while since there was no way to distinguish, but users kept using v2, which was failing cryptically.

So now we need to try v1, and if it fails try v2. But that means that our monitoring on errors are all fucky now since half the errors arent real errors.

In the grand scheme of things I guess it’s a relatively small design issue, but it sitting right in the middle of their one and only critical path, which makes it really painful.

I’ve alrwady contacted support about it and their response has been, more or less, “oh, uhhh... huh.”

</rant>


One of the nice little features that I appreciate about Twilio, and have adopted myself, is their API keys are always prefixed with a 2 character type.

e.g.

"account_sid": "ACXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX", "phone_number_sid": "PNXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX", "sid": "CAXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX",


Hard-won pro-tip: IDs should pretty much always have some sort of prefix when stored as strings, even if you swear you're never going to change them. You're not going to want to try to divine which id was which based on string length or what chars are where later.

I qualify "when stored as strings" because id numbers in database rows are in a sufficiently strong context on their own, for instance. But once they get serialized to strings, you really ought to stick a prefix on them. You may only change 10% of your ids, but the saving on that 10% will be well worth the cost of prefixing your ids.


But why this doesn't needed if you use numerical ids?

If you work on a sufficiently long lived system, you also discover that you usually want string ids instead of numerical ids.

Among other things, numerical ids are subject to enumeration attack (especially if exposed in a GET url parameters).


Yeah, we do this too in Arvados (https://arvados.org). All our uuids consist of three parts separated by a hyphen. The first part denotes the cluster, the middle part denotes the type of object, and the last part is a random string. E.g.

  qr1hi-j7d0g-a68mg4e5oh37oek
where 'qr1hi' is the cluster identifier, 'j7d0g' means 'project', and 'a68mg4e5oh37oek' is a unique identifier.

Do the same calls to the same APIs from the same customers ever change from v1 to v2 (or vice versa)? If not, can you record if it was v1/2 and then always make the same call after that? And then don't record errors for your initial probe request? Then your data will be clean and you'll make a lot less calls.

Yeah, we have some caching and retain metadata to avoid making more calls than we need to.

It’s just frustrating that we need to keep this information around on our side to mitigate issues with the API that is central to their entire business.


What's the big deal in checking if each version works?

Sounds like you just need to ignore the errors you get when testing the API version, e.g. wrap a try/catch block around the request.

If that's your big gripe with the API, you have it pretty good!


if you've ever logged into pager duty and tried to use it you'd know UX/ergonomics is not their strong suit.

You do realize the v1 API was deprecated last year?

That's only true of the REST API. The Events API has it's own v1/v2 split and I don't believe there has been any mention of the v1 Events API being deprecated.

that's it, this comment is a nail in the coffin of their future IPO.

They just raised a round in September of 2018 at $1.3B so it will be interesting to see where the public market values the company.

2017 revenue: $79.6MM 2018 revenue (est): $117.8MM 2019 revenue (est): $164.9MM 2020 revenue (est): $214.3MM

Safe guess potentially 10x current year (2019) which would be $1.6B.

Considering that they need an IPO pop the shares would be priced about 20-25% lower, bringing them back to that $1.3B private round valuation.

Unless the bankers are very optimistic and it will trade higher in which case they would be looking to price at the $1.6-1.7B range and hope that it pops to $2B at open.


You need to dream bigger. Palantir looking to go public at 45X revenues.

https://news.crunchbase.com/news/palantirs-2018-revenue-said...


That would be so absurd

I think the market isn’t crazy so they can start at whatever and feel reality will slap them sober. Like Dropbox and Snapchat experienced.

But.. they're a development agency! Right?

2B would be a reasonable price. Who knows, they may get acquired by the tech behemoths.

>In October 2018, the Company entered into a cloud services agreement for third-party hosting of the Company’s software platform through October 2021. The Company is committed to spend a total of $13.5 million for the services through the term of the agreement

Not as bad as some of the others lately but this works out to around 375k/month.


Congrats on PagerDuty!

I set my PagerDuty to the meow sound and I often refer to our processes as meow driven development as in I don't like my phone to meow desperately.

If I worked in the same office as you I would 100% play Jingle Cats when walking past.

If I don't change my page sound frequently enough I start getting a weird Pavlovian response.


> If I don't change my page sound frequently enough I start getting a weird Pavlovian response.

I had to do the same with my alarm clock, and eventually settled on talk radio because it didn't seem to cause the same effect.


When I was a kid I got a two deck cassette player and alarm clock combo. I used it as my alarm for about 16 years, and I still get a weird feeling if I hear that precise tone. It makes me feel a bit sleepy and disoriented for a split second. With the same clock I also started to wake up to the sound of the system powering up the speakers to play the torn before the tone sounded. It seems I’d slap the snooze button in my sleep so I had to get a new alarm.

We get paged extremely rarely. The last one was Nov 23. So no weird reflex here.

I use the second photo from here[1] as the background on the caller when I get paged.

[1]: https://grist.org/climate-energy/these-new-high-res-photos-o...


The feature we need to add to https://amixr.io . No idea how exactly but we need to...

Why would any sane person buy shares with this power vested in the board

"authorize our board of directors to issue, without further action by the stockholders, shares of undesignated preferred stock with terms, rights, and preferences determined by our board of directors that may be senior to our common stock;"


I believe this is how it typically works. The board is hired by the shareholders to act in their interest. The board has a lot of powers. The shareholder if unhappy with the boards performance can fire the board.

Did you not read the bit without further reference to share holders - the board can issue new shares to dilute existing share holders.

This language describes "blank check" preferred stock -- which about 98% of IPO companies have in their governing docs. So, while you might not agree with the philosophy behind it, if it is a veto item on investing, you will have a pretty limited portfolio. See page 12 of this IPO study

https://alerts.davispolk.com/10/3818/uploads/2018-non-contro...


Id rather stick to shares where shareholders have to agree to disallowing pre-emption rights.



Applications are open for YC Summer 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: