Goodbye SQLCruise, Hello Tech Outbound.

Blog Posts

It’s hard to believe that it’s been seven years since I had an idea, booked a cruise, and emailed a few friends:

SQLCruise: My Next Bad Idea

SQLCruise: My Next Bad Idea

I really didn’t expect anybody else to join in with me – there was a ton of risk, and there hadn’t been anything like that in the SQL Server community before. Other communities were doing it, though: Burleson offered Oracle classes at sea, and a company called Geek Cruises (now defunct, archived PDF brochure) had done something similar with geek superstars like Wozniak and Linux Torvalds. I figured my buddies would poke holes in the idea, and that’d be the end of it.

They all seemed to think it was gonna work, but Tim Ford ( was the only one crazy enough to join in. It looked easy in the rear view mirror, but I totally don’t blame Andy, Jeremiah, and Tom for not joining me in starting the company – there was a ton of risk, and success was by no means assured.

Amazingly, the first voyage was a success.

SQLCruise #1

When I look back on photos of the first SQLCruise, I break out into giggles. I met so many people who later became MVPs, went to work for Microsoft, became consultants, and are giants in their fields. It was even where I met Kendra for the first time, picking them her from Palm Beach due to a travel mixup.

SQLCruise was a total success in the metric that mattered to us: not money (I don’t think we even broke even because of all the prizes we gave out), but building relationships in the community. It was all about meeting other fun-loving community members, talking about databases, and hanging out at sea. As the photos & tweets filtered back to land, it was obvious from the outside that it was a great formula.

We ended up doing another, then another. Tim wasn’t just “a speaker” – he and his wife Amy quickly became the driving forces behind the event. They’d clearly found a calling.

Tim & Amy took it over, growing and changing it.

Within about a year and a half, I bowed out to focus on growing the consulting business, and Tim and his wife Amy took SQLCruise over and ran with it. They not only made the event their own, but they took it to the next level. When you look at my 2010 email, and compare it to what it’s become today, it’s clear how much work they’ve put into it to make it a serious event. Ain’t no free casual tweetup anymore.

By charging admission and taking on sponsors, Tim & Amy were able to add more guest instructors. The 2017 cruises had a total of 8 instructors, including 2 Microsofties:

SQLCruise 2017 Technical Leads

Damn, that’s a list of smart people.

This week, they’re on a boat.
You should be with ’em next year.

You might be seeing the #SQLCruise tweets in your timeline this week – they’re on an Alaska cruise this week. (Funny timing, it’s the same week as our company retreat, but we’re on a different boat.)

I’m not associated with SQLCruise at all anymore, but I still believe it’s one of the best investments you can make for your career. The outside-of-class interaction time between attendees and instructors means you can accelerate your career and build valuable relationships. You get a ton of valuable time with some of the sharpest people in the industry – people who can help you make connections you wouldn’t able to make otherwise.

And now, the Fords are taking it to the next level again, renaming it as Tech Outbound. They’ve already announced their 2018 cruises, and they’re talking about doing events above and beyond cruise ships. I’m so excited to see what they come up with.

If you love data and you want to meet other people who are just as passionate about it as you are, but also love getting out and seeing the world, check out their upcoming events. Life is short – get started building relationships that you’ll cherish for years.

Read the Comments

The DBA Job I Turned Down – #TSQL2sday

Blog Posts

For this month’s T-SQL Tuesday, Kendra Little asked us to share our interview patterns, anti-patterns, highs, and lows.

Several years ago, I was talking to a prospective employer via phone, and things were sounding really interesting. I made it past the first couple rounds of remote interviews, and they wanted me to come into the office for the final round.

The interview started with a tour of the offices, and things seemed alright enough at first until we got to the part of the building where the technical people worked.

We stepped into one room maybe 20′ x 20′, and it was completely packed with desks, classroom-style –  and 15-16 employees. Each desk was maybe 3′ across – wide enough for a single monitor, a keyboard, and a mouse, and that’s it.

The desks were all next to each other, and I mean next to each other. Without leaning, each person could reach out their arms out sideways and tap both of their neighbors on the shoulder. Furthermore, each person could turn 90 degrees, and tap the monitors of the row ahead & behind them. I’m not even sure how one person could have taken a bathroom break without asking other people to move. Maybe they had to go at designated times. Maybe they weren’t allowed to go at all.

Dramatic recreation

As if on cue, the entire room turned to look at us, then turned right back to their monitors to continue working.

This was one of those times in my life where I really wish I had a photograph of my reaction. It must have been horrific, because as we walked to the next room, my guide said, “No no, don’t worry, you wouldn’t be working in that room. You’ll be in another area.”

That was it, though – I was shellshocked and done. I smiled and nodded my way through the rest of the interview.

Shortly thereafter, the recruiter called me back with the great news that I’d been offered the job. I turned it down. Sure, that wasn’t my office today, but if the company would do that to any of their team members, there was nothing stopping ’em from doing it to me later.

Read the Comments

Thoughts on Building Community Tools

Blog Posts
No Comments

The nice folks over at the SQL Data Partners Podcast, Carlos & Steve, talked to me and Chrissy LeMaire (of fame) recently about what it’s like to build an open source project with the community.

Here’s the episode: Building Community Tools – and you can listen to it while you read this. I really like how they edited it together, mixing their thoughts in with ours.

Since the podcast, I’ve had a few other thoughts that I wanted to include.

At first, you have to be consumer #1.

Build something that solves a real pain for you every day. If you’re feeling the pain, other people are feeling it too, and there’s nobody else who understands your pain the way you do.

If you use your tool every day, then you’ll be motivated to keep making it better. You’ll even get excited when other people propose improvements because their work will make your life even easier. (Eventually.)

On the other hand, if you build something thinking “this would be cool for somebody,” then you’re already at a disadvantage because you don’t really understand the target market.

Sometimes, you’ll get the wrong consumers.

People will see your tool and say, “Wow! That’s cool, but it’d be even better if it did ___.” They’ll ask you to write code to do ___, and you’ll feel guilty because they like part of your tool, but you want them to like the whole thing. Later, people will even say things like, “Your tool just sucks because it doesn’t do ___,” or “Your tool doesn’t do ___ the right way.”

You have to be comfortable saying no.

I get this a lot with sp_Blitz – people say things like it should do security audits, or run on Azure SQL DB, or be written entirely in PowerShell. Those are all great ideas – but they’re for someone else. I wouldn’t be a consumer for that, so I’m not really the right author to write those tools. They’re welcome to write ’em, though.

Later, you’ll get the wrong developers, too.

At first, when you start getting code contributions, you’ll be just so excited no matter how bad the code is or no matter what features it tries to accomplish.

Over time, that gets old because you are the one who has to support it. Someone will do a drive-by contribution, put in really bad code, and then not be around to answer the inevitable bug reports.

To deal with that, read Daniel Bachhuber’s wonderful post, My condolences, you’re now the maintainer of a popular open source project. It does a great job of capturing the highs and lows of building something that captures public attention. You end up feeling guilty because you can’t do enough work – there’s always people wanting you to do even more. You have to set boundaries, make yourself happy first, and do what you can, where you can.

Building open source costs real time and money.

At first, since you’re consumer #1, the tool you’re building saves you time and money.

Later, as you start dealing with other consumers and developers, a open source project can become less about solving your own problems, and more about giving back to the community. Our company uses sp_Blitz/sp_BlitzCache/sp_BlitzIndex/etc in our daily work to put bread on our table, but there’s a cost to sharing it with the public. No, I’m not talking about the cost of competition – we don’t lose money if other consultants use our tools to do health checks – but we have to maintain & document the code for other folks to use it.

Right now, we spend between 2-4 days per month doing coding, testing, documentation, and releases. That equates to one SQL Critical Care of lost revenue per month – so we lose around $85k/year on the First Responder Kit. You have to think of it as an investment, and be conscious of what you’re buying with that. Open source projects can be of business value, marketing value, or just plain giving back to the community – but you have to make a conscious decision as the project grows to make sure you’re getting the value you need out of the project.

Read the Comments
Porsche 911 Targa that I definitely don't own

Should You Take a Salary When Starting a Consulting Company?

In the startup world, the general advice is for founders to pay themselves $50k-$100k/year until they get outside funding. Your net worth ends up higher – but only if you hustle and grow the value of your company, since your net work mostly consists of your ownership stake.

When you build a consulting company, there’s an unbelievable amount of unpaid labor to be done: hiring staff, training them, managing them, improving the deliverable product, marketing, sales, contract negotiations, expenses, accounting, taxes, web site maintenance, you name it. It’s all too easy to say, “We should really pay ourselves for doing this unglamorous work.”

But when we started Brent Ozar Unlimited, we didn’t take any salaries at all. We figured that since Jeremiah, Kendra, and I were all consultants, we could make money directly. Each partner got a percentage of the consulting work that they personally did, but the rest of the money stayed in the business.

Porsche 911 Targa that I definitely don't own

Porsche 911 Targa that I definitely don’t own

We figured that if we were going to hire people, we needed to build a little money fort inside the business. In the event that an employee didn’t have revenue associated for a few months, we still wanted to be able to pay their salary. (That goal ended up coming in handy, and I’m really thankful – if it wasn’t for that, we’d have probably had to close the company’s doors.)

When we finally had enough unpaid labor involved (due to growing head counts, taxes in different states, and selling goods online to EU citizens), rather than paying ourselves for it, we hired an admin to offload 40 hours a week of it – but the unpaid work for the partners continued (and continues!) to grow. It’s just a byproduct of running a growing company.

Taxes make this tricky.

Here comes the really challenging part: in a pass-through LLC, the founders get taxed on profits even if the profits stay in the business. To keep things overly simple, let’s say in a one-founder bootstrapped consulting company:

  • Brings in $150K of consulting revenue
  • The founder only takes out $100K of that, leaving $50K in the business
  • At the end of the year, the business had $20K of expenses, and $30K left in checking
  • The founder gets taxed on $130K of revenue, even though take-home was only $100K

The more successful you are, and the more you want to sock away in the business’s savings, the more you get penalized on your personal taxes. Businesses are encouraged to cut things as close as possible, keeping as little cash as possible in savings.

SaaS startups don’t have this problem because they don’t make any significant income for years. Their cash comes from investors, so it’s not revenue.

So should consultant founders take a salary?

The Porsche 911 Targa I do actually own

There’s probably a little survivorship bias here, but in our case, I’m really glad we chose not to take salaries. (And I still don’t take one.) Cash in the business’s checking account doesn’t let me buy that Porsche 911 Targa that I want today, but it fuels the company’s fire to hopefully let me get one years from now.

In your case, talk to a good attorney and a good accountant. When you’re starting a business, ask around in your circles for local folks who specialize in this kind of thing. Attorneys & accountants will usually have a 30-minute call with you for free, and a good one will ask you questions about your goals, your future tactical plans, and what success looks like to you. The questions they ask will enlighten you and help you decide.

Read the Comments

Dream Car Garage: 4 Cars, $250k

In the last couple of posts in my dream car garage series, I’ve kept a low budget. Screw that – let’s spend some money. Let’s pretend we’ve hit the big time, maybe sold the company for a few million bucks, and it’s time to build a really classic garage to see me through retirement.

1987 Porsche 911 Turbo ($73k)

Guards red, my favorite 80s 911 color

These did 0-60mph in just under 5 seconds, which was unbelievable at the time – a total handful of a car. 911s have long been infamous for their snap oversteer (video), and un-intuitively, the fix was to bury the throttle. It takes a lot of guts to stand on the gas when the car starts to go out of control, and as a result, there’s less and less original (non-smashed) 911 Turbos out there.

I had many a 911 poster on my wall growing up, and while it was about the (perceived) speed then, it’s about the looks today. What a beautiful, iconic shape. I love the Guards Red, the Fuchs wheels, the whale-tail spoiler, even the grinning American bumpers.

There is no runner-up here. A red 911 Turbo has to be in this collection. The end.

1986 Ferrari 328 GTS ($50k)

The dream car garage is obviously in Hawaii

Can’t you just hear the opening theme to Magnum, PI? I didn’t watch the show at all growing up, but this car’s image is forever attached to Tom Selleck. I bet the hardest part about owning this car isn’t the high repair bills – it’s the challenges of mustache inadequacy.

Putting this list together, I was really torn about not getting a Porsche 911 Targa – I’m a huuuuge fan of the targa top – but finding this gem reminded me that there’s more than one brand that builds a beautiful targa top. I know it’s not fast or aerodynamic, but my goodness, it’s good-looking.

Runner-up: 1970 Porsche 911 Targa. Yes, I’d put two 911s in the garage, but this one is a little tricky. I love Targas, but the only reason this particular one is even remotely affordable is that there’s almost nothing original left on it (different engine, paint color, spoiler, etc.) I’d buy it just to enjoy it – I’m not a purist. I’d wanna get the ducktail spoiler off there, though – I love ducktails, but they don’t make any visual sense on a Targa.

1973 Jaguar XKE V12 Roadster ($61k)


Enzo Ferrari called it the most beautiful car ever made. I wouldn’t quite go that far, but it’s one hell of a looker. This car’s heritage is what makes me really want an XKR-S, its modern grandchild. (That and the addictive exhaust note.)

I didn’t realize until I was writing this up that I’d managed to pick three red cars in a row. I’m partially red-green colorblind, and Erika has already told me that I’m not allowed to buy any more red cars. Maybe I’m overcompensating here. All three of these cars are just smashing in red, though.

This one in particular is an automatic, which keeps the price down. That’s completely fine with me – the slushbox keeps the price down as opposed to the utterly crazy Series 1 prices. As crazy as it sounds, even in the quarter-million-dollar collection range, you have to be a bit of a price-shopper to get yourself into a Jaaaag.

Runner-up: 1967 Mercedes-Benz 250SL. I love, love, love W113 Pagodas. They were among the first collector cars that I showed Erika. She responded, “It looks like it’s owned by someone who couldn’t afford a current Mercedes.” I love her anyway.

2007 Audi S8 ($39k)

Me? Going 150mph? No, Officer, must have been someone else.

No, it’s not an all-red garage.

Looking at the above list, there’s nothing I can use to take some friends out to dinner. None of the above three are really practical for, say, a trip to the airport with luggage. So for the last car, I went for a simple, practical sedan. Nothing to see here.

But when you turn the key, there’s something to hear: the monster V-10 engine from Audi’s corporate brother Lamborghini.

This Audi’s shape has aged really well, and I think it’ll continue to look great in the decades to come. Sure, it won’t stand out – that’s what a sleeper is all about – but it’ll be a joy to drive while it sneaks under the radar.

Runner-up: the 1991 Jeep Grand Wagoneer would make a fun (if unreliable) daily driver. Holy cow, those things have gotten expensive. I’d also love to think I’d do the 2008 BMW M5 with a stick shift and a V-10, but depending on the traffic where I’d end up living, a stick might be a bit more of a hassle than I’d want in a daily driver.

Update 2017/07/18Daniel Janik blogged his picks too.

Read the Comments

What if Microsoft Open Sourced Books Online?

Careful observers may have noticed a few little icons at the top of Books Online pages recently:

Click to see Books Online, because I know you can’t get enough of that

The little round contributors icons take you to their Github profiles, like Rick Byham’s.

The all link – well, that doesn’t actually work, because it goes here:

Which isn’t visible to the public.


But note that not only is .NET open source, but even Azure’s documentation is open source. That’s right – you can make a pull request and update Azure’s equivalent of Books Online.

So it’s entirely conceivable – nay, I say probable – that SQL Server’s Books Online will end up accepting pull requests from mere mortals like you and me. It won’t save Microsoft money by any means – imagine the junk “contributions” they’re going to get – but in the age of running SQL Server on Linux, they’re going to need all the help they can get. The Linux community already understands how open source works, and they’re going to expect to be able to give back to the documentation efforts as folks learn what it takes to run SQL Server successfully on Linux.

It’s a brave new world. I love it.

Read the Comments

Updating My About-Me Slide, 2017 Edition

As I was prepping for SQLSaturday Houston, I realized it’s been two years since I updated my About-Me slide at the start of my presentations. Here’s what it was for quite a while:

About Me, circa now

About Me, circa 2016

This time around, I decided to start from scratch and ask:

Who am I writing this for? Like any presenter, it’s tough for me to generalize everyone in the audience into a single bucket. Some folks have chosen to sit in my session because they recognized my name, but others don’t know me or why I’m talking about this topic. That makes defining a single target audience pretty difficult, so I’ll define 2 attendees:

  • The first-timer: doesn’t recognize my name, and doesn’t know what my job/blog/SQL history is
  • The regular: knows my name, has seen me present before, read my stuff, and knows what they’re in for

What do I want them to learn on the slide? I want both types of target attendees to learn something, so I decided to split the slide into two halves:

  • For the first-timer: my job history
  • For the regular: my community projects

Business on the left, party on the right

I’ll spend 15-30 seconds covering each side by saying:

Left side: “My name is Brent Ozar. I make SQL Servers faster and more reliable. I got my start as a developer, then moved into database administration, and VM and SAN administration. I’m talking about today’s topic because (insert relevant detail here, pointing at the part of my job history where I needed to work with that technology.) If you need me, the best way to contact me is”

Right side: “If you like today’s session, you might like some of my other free community projects. is an online conference where you pick the sessions, and it’s all online for free. sp_Blitz is my free SQL Server health check, and I give away a ton of other stuff at these other sites. How many of you have used something from this screen? Okay, great! Now let’s get started.”

It’s more text-heavy than I’d prefer. I’m tempted to break it into two slides – getting paid and giving back – but that feels like it’d be violating the one-page-resume rule. I’ll probably cut down some of the right hand list to just my favorite community contributions though.

Read the Comments

The Freelance Calculator

If you’re thinking about going freelance, here’s a handy Freelance Calculator by Chelsea Shaw (@chelshaw).

See the Pen Freelance Calculator by Chelsea (@chelshaw) on CodePen.

(Big props to Chelsea – I originally just sketched out the UI, and she knocked out all the code! You readers are amazing.)

  • Hours per week = the number of hours you’re going to work, but work includes both billable hours as well as non-billable (paperwork, email, travel, blogging, presenting.)
  • % billable = the percent of your work that you can actually charge for.
  • Hourly rate = what the clients actually pay.

Being an independent means you can change any of these things, whenever you want. (And you will, through the course of your career.) In a perfect world, you’d track these 3 numbers over time to know if things are getting better or worse, and track the success of your experiments in passive income (training courses, plugins, SaaS apps, etc.)

The equation:
Hours worked * (% billable) * hourly rate = your revenue.

But don’t think that the last number is the only important one.

Sometimes you want to work less.

Spending time with your family and friends is awesome, and wouldn’t it be nice to retire, too? You need money to do those things – so your rates need to be going up, not down, to maintain your standard of living.

To work less, you have two options.

1. Short term: do less unbilled work.

You don’t get paid (directly) to:

  • Do marketing to bring in new leads
  • Convert those leads to buyers (do sales)
  • Do support to keep those buyers happy
  • Sharpen your skills via learning, experimenting, building
  • Build passive income tools like training classes or apps

But you get paid indirectly (later) for those investments. If your calculator produces less than 10-20 hours of unpaid work per week, you’re probably underestimating the above work requirements.

You can coast on your existing network & skills for a while, but do that for a year or two, and you’ll suddenly find yourself in trouble. This is how freelancers end up taking a gig they don’t really want, just out of desperation, or end up joining a consultancy.

Don’t want to do unpaid work? You can pay someone else to do it. Sometimes they do it better than you could possibly do it yourself (like my accountant or my attorney or my coworkers.) Sometimes they don’t – like when I’ve tried using Google or Facebook advertising to bring in leads. Compared to what I get from the blog & presentations, the lead quality hasn’t been as good, and it’s been expensive as hell. But if you hate marketing or content generation, you can buy your way out of it. It’s just a matter of setting the right settings on the calculator – your revenue may need to go up in order to justify paying someone else to do stuff.

2. Long term, you need to raise your rates.

Raising your rates is the key to having more flexibility to pick between paid work, family time, relaxation, or investing in your future. When you raise your rates:

  • The less billable hours you have to put in for the same revenue, and
  • The more time you can spend on marketing or building passive income products like training courses, which
  • Increases the demand for your services, which
  • Means you can raise your billable rates again

For more info on that, check out Brennan Dunn’s Double Your Freelancing. (That’s not an affiliate link – I just really believe in those concepts, and I think the more people who know about that, the better.)

Optimize your calculator settings over time.

Over time, you should be optimizing this calculator to find the right work/life mix for you, and optimizing your services & rates to minimize the amount of work required. These kinds of conscious planning decisions are the difference between a okay freelance/consulting career and a great one – hopefully short, too, so you can retire and enjoy time with your friends and family while you’re still able.

Read the Comments

What #SQLSaturday Organizers and Volunteers Do

Last weekend, I had the privilege of speaking at SQLSaturday Houston. Allen Kinsel, Devon Leann, Joe Hellsten, and so many more volunteers did a fantastic job of putting on a totally free event for hundreds of attendees.

I’ve never organized one of these events myself – I’m just not that ambitious – but because I get to see some of what goes on behind the scenes on event day, I want to give you a quick peek at just how much selfless work these volunteers do for you, dear reader.

Here’s just a sampling of the work they have to do before the event:

  • Find a venue – which is surprisingly hard to find one space with multiple classrooms, projectors, WiFi, and keep costs as low as possible
  • Announce a call for speakers
  • Line up sponsors – companies who are willing to fork out money to pay for your free training
  • Recruit volunteers who are willing to work leading up to the event, plus on the day of the event, and may not be able to even see the sessions they wanna see
  • Pick the sessions, trying to curate a well-rounded agenda and foster local speakers
  • Plan coffee and hydration for attendees, get a lunch caterer
  • Get bags for attendees, and coordinate sponsors who want to put swag in attendee bags
  • Deal with accounting/banking/paperwork for all this, trying not to lose money or go into personal debt for cost overruns
  • Email speakers, sponsors, and attendees leading up to the event so they have the right travel logistics
  • Decide whether to do online speaker feedback (which attendees rarely use) or print the feedback forms (which attendees use and speakers love, but takes more work)
  • Help vendors ship their booth gear to the closest location, even to the volunteers’ houses and the volunteers drag the stuff into the event
  • Stuff hundreds of attendee bags full of sponsor swag
  • If there’s extra money, maybe do a speaker dinner or drinks

Allen Kinsel finishing up SQLSaturday Houston, getting ready to start the raffle

And then on the day of the event:

  • Bring in the coffee, breakfast, water, snacks
  • Coordinate with sponsors to set up tables and raffle entries, make sure the sponsors are happy campers and will come back next year
  • Put signage around the venue so attendees know where to park (remember, hundreds of people)
  • Put signage inside the venue showing which sessions are where
  • Set up the speaker room and make sure all the speakers showed up
  • Check in attendees, helping some folks print their entries, give them their bags
  • Coordinate volunteer room monitors
  • Help speakers who didn’t bring the right cables, or who have projector problems
  • Find speakers who can cover for no-show speakers
  • Help the lunch caterer set up and tear down
  • Run a raffle drawing to help promote the sponsors
  • Clean up the venue after it finishes
  • If there’s extra money, maybe do an attendee party afterwards

If you’ve attended SQLSaturday, and if you got value out of it, you should think about volunteering to help put the next one on. SQLSaturday runs on the backs of regular people just like you. Speaking isn’t the only way to help – and in fact, I think the speakers might just have the easiest job at the conference.

Wanna help? Find a nearby upcoming SQLSaturday. The event page will have a call for volunteers. You should volunteer. You’ll be exhausted afterwards, but incredibly fulfilled. These people are just like you, and they make a huge difference in the SQL Server community.

They are the SQL Server community.

Read the Comments

Book Review: Google’s Site Reliability Engineering #tsql2sday

For this month’s T-SQL Tuesday, Grant Fritchey’s topic is Databases and DevOps.

Summary: you should skim this free online book to see inspiring ideas of how administration works at scale, although don’t expect to put the practices into place without management buy-in.

Let’s get one thing out of the way first:
you, dear reader, probably don’t work at Google scale.

Google faces similar problems that your employer does, but just at a different quantity. Instead of keeping a 3-server applications reliable, they need to keep 3,000-server applications reliable. As a result, they have very different budgets than you do, and it gives them the luxury to treat it as a serious discipline.

Free Site Reliability Engineering book

In their free Site Reliability Engineering book, they share some of the lessons they learned about:

  • Designing service level objectives
  • Tactics like deployments, monitoring, automation, and release engineering
  • How to load balance and handle overload, and much more

You really don’t need to read the whole book – just skim it, and you’ll take away interesting concepts and stories. While DevOps and SRE aren’t the same thing, you’ll start to see how your DBA duties, DevOps duties, and developer duties all blend together to work towards the same business goals.

My favorite concept: error budgets

Say you’re given a 99.5% uptime goal. Instead of thinking of it in terms of time, think of it as, “0.5% of my service’s requests may result in errors.” Maybe it’s the entire service is unusable, maybe it returns a failure of some kind, maybe it times out.

Instead of aiming for 0% errors, aim for 0.5% or less errors.

0.5% is your error budget, and you’re expected to spend it.

You may spend it on planned outages for software deployments, spend it accidentally in the form of unplanned outages, or purposely for things like patching or major app code deployments, or maybe just plain old experimenting by cutting costs. This starts to set the stage for why we need DevOps – developers want to spend part of the error budget on deployments.

DBAs usually aim for zero errors. DBAs don’t want to spend the error budget at all, but the business needs us to. If you don’t use any of your error budget, that’s a problem because it indicates that you’re probably spending too much (money or resources), not doing the right upgrades/patching to keep your application current, or trying to make developers work too hard to build absolutely perfect deployments every time (which cost a ton of money to build). You should probably look at ways you could cut costs in order to get closer to the business’s objectives.

Prefer videos? They’ve got videos too.

Check out the reliability engineering talk from Google Cloud Next 2017:

I’m certainly not saying that Google does everything right, and that you should model all of your practices after theirs. That’s ludicrous – they’re huge, and they have huge budgets. But there’s some interesting lessons about your own database operations and deployments.

Read the Comments