Mangoteque

After decades in Silicon Valley helping companies get better at software delivery as a sysAdmin, tech lead, IT director, architect, and enginering leader, I set out to help many more engineers live better lives.  No more working nights and weekends, no more"death marches" and "war rooms" to deliver, no more humans doing repetitive work computers can do.  When engineers like their jobs they produce better results, which allows executives to reach their goals, and investors to deliver fabulous returns. Win-win-win.

Leveraging my lengthy experience in the DevOps movement, I've helped *dozens* of companies achieve their investment goals without bringing in contractors or installing chatty software with no context.  

As Deming said in Out of the Crisis: "I remind the reader that the improvements took place with the same people and no new equipment."

Many companies try to "work harder" their way out of their challenges.  With Assessments and engagements, CTOs and I partner to get organizations working faster and better, so CTOs can go to the board with total confidence, and the board knows they will get the best results.  Win-win-win.

Founder & CEO

Helpful Engineering

Leading and organizing the global "DevOps Team" of a few dozen active engineers supporting a large global organization of technologists actively fighting the COVID-19 pandemic.

Global DevOps Team Leader (pro-bono)

SolarWinds

•Recruited by Librato CTO to join team preparing for acquisition by SolarWinds. Accepted newly created role after acquisition to relieve Director of Engineering of increasing responsibilities. 
•Charged initially with one nationally dispersed team. Gained additional national/global operations teams as company continued to acquire companies, integrating products/services and transitioning to new platform. 
•Developed Global Site Reliability Engineering (SRE) team as direct reports grew, managing 19 individuals across five teams to drive infrastructure, software, and network evolution and support for ~1M customers.
•Built shared platform for SRE team. Leveraged automation, consistent tooling, reuse of code, and rapid deployment of new services. 
•Managed $5M CapEx/OpEx budget, driving innovation, capability, and capacity across data centers.
•Developed global strategy and comprehensive roadmap to enhance activities across functions; collaborated with executive leaders to assess needs and deliver technology solutions.
•Accepted numerous speaking engagements regarding automation, digital transformation, and improvement possibilities effective SRE teams can provide.  

Select Achievements
•Ensured superior products and customer satisfaction by adopting agile incremental work delivery methodology; deployed 1K+ per year in contrast to typical company deployment of twice per year. 
•Migrated Librato from Amazon Classic to Amazon VPC network environments while migrating services with almost zero downtime; resulted in reduced costs, faster recovery from failure, and greater agility.
•Led digital transformation for five acquired teams: 
--Transitioned Papertrail to new data centers and AWS, doubling capacity while holding costs constant. 
--Transitioned Pingdom to AWS using containers as primary deployment mechanism.
--Reduced alert load 66% for acquired Loggly team, allowing additional time for infrastructure improvement with no extra downtime.

Global Head of Site Reliability Engineering

salesforce.com

•	Served as Architect in technical operations, working with individual contributors and executives to improve efficiency, speed, reliability, quality, and effectiveness across organization.
•	Transformed company mindset to embrace automation and cross-functional collaboration reflected in strong DevOps methodology. Served as keynote speaker/SME internally & externally. 

Select Projects
•	Served as sole architect on first Infrastructure as Code team focused on servers; established consensus within executive and technical leadership on this major global operational change.
•	Contributed to first six-member Infrastructure Orchestration team to set up deployment code to work in concert with new Infrastructure as Code paradigm.
•	Rescued project focused on creating new internal DNS as a service; wrote project scope and worked with senior software and network engineers to deliver environment and infrastructure. Transformed internal resources, automated configuration of hardware and software, to allow for 24/7 agility.

Architect, Infrastructure Engineering

Tagged

•	Revamped operational infrastructure (~5B page view per month); designed and built complete production architecture transitioning from bootstrap focused methodology to cloud principles. 
•	Created and led five-member Blog Team and 15-member DevOps@Tagged discussion group. 
•	Served as main point of contact for Engineering Infrastructure team focused on delivering new capabilities.
•	Mentored three interns and junior staff on monitoring and infrastructure automation.

* Presented "Enterprise Real-Time Visualization with Graphite" at Bay Area Large Scale Production Engineering Meetup (6/2011)
* Presented "The sFlow Standard: Scalable, Unified Monitoring of Networks, Systems and Applications" at Velocity 2012 (6/2012) 
* Contributor "Monitoring with Ganglia" (O'Reilly) pub 11/2012

* Completed "Leadership Training' class facilitated by MyBrainSolutions.com
* Completed "Crucial Conversations" communication class

Senior Systems Engineer

Terracotta Inc.

•	Responsible for complete production and internal architecture of leading-edge Java clustering startup; managed two direct reports, one remotely, while reporting directly to CEO.

CTO SWAT Team

DHAP Digital

• Worked on internal and client effort equipment for Toyota web farm; performed architecture, administration, and analysis duties.

Sr. Systems Administrator

Cable & Wireless LLC

• Served as technical lead for systems administration group responsible for global content delivery network of 1300+ pieces of network attached equipment.

Lead Systems Adminstrator

Epylon Corporation

• Responsible for architecture and implementation details of a complete production, as well as the development of e-commerce installation.

Lead Systems Engineer

Miller Freeman Inc.

Senior Systems Administrator

After an investment from Private Equity, portco CTOs are under immense pressure to deliver more.  More often and with more quality, but more often than not, without more people.  But how?

You're a great CTO, but you don't have the time to coach everyone. It's too much.

◆ Customers are complaining about stability.
◆ Ops wants people to open more tickets.
◆ You *wish* the Dev teams were doing it wrong and shipping at least every two weeks.
◆ The board wants better EBITDA.
◆ You have separate DevOps, CloudOps, and SRE teams and it has to make sense, now.
◆ The CEO wants a plan for the next year, now.
◆ You've done 3 acquisitions and have 8 data stores.

With a big investment comes the expectation for big results. EVERYONE wants more.

As a long time DevOps and Silicon Valley veteran, I've improved operational excellence working with *dozens* of companies.

☆ speed up the Ops team
☆ streamline delivery
☆ increase resilience
☆ ship more often
☆ improve product quality
☆ enable engineers to innovate
☆ make the CISO smile!

My relationship with your engineers and leaders is more mentor than consultant.  I've faced the same problems at many different scales.  Applying the science behind the DevOps movement helps us improve development, operations, quality engineering, resilience, deployment, security, compliance, and more.

I teach engineering leadership daily how to make real changes, something CTOs simply don't have the time to do.

We know that companies who are elite at software delivery are “twice as likely to meet or exceed their organization’s performance goals”.   That's why I designed the DORA-based 2 Week SDA, so your teams have a specific playbook on how to become elite at software delivery.

We're a good match if you want to leverage deep experience and deliver on expectations, without everyone hating their jobs.

► “Through a decade of work with Thoma Bravo portfolio companies, Dave has taught our product teams how to dramatically improve business outcomes and to discover where time comes from.”  - Paul Zuber, Operating Partner and Head of the Technology Group

► “Few consultants have the ability to help you get where you want technologically as well as to truly improve your team itself.” - Nathaniel Barnes, CTO, MeridianLink

🏃‍♂️ DM to go into the next board meeting confident about the plan
💡 Learn more at mangoteque.com

Years ago, I used to care whether a team was called the "DevOps Team".  DevOps is a philosophy, not a job title.

But I've stopped dying on that hill.  James Ausman asked what my beef is with talking about "The DevOps Team". 

Thanks to Gemini, 10 team names that you would probably never say:

1. The "Write Tests First" Team
2. The "Release Often" Team
3. The "Get Feedback Quickly" Team
4. The "Keep Things Simple" Team
5. The "No New Bugs" Team
6. The "Don't Break Stuff" Team
7. The "Readability" Team
8. The "Measure Everything" Team
9. The "Refactor Old Code" Team
10. The "Ship an MVP" Team

Almost every single one is part of the DevOps philosophy so it would be the responsibility of "The DevOps Team".  Everyone else gets a free pass! 😂 

Unless I know that you know better, I'm over it.

What doozies did Gemini miss?

Dave Mangot

I partner with PE portco CTOs to deliver on the investment thesis by changing the culture of their existing org from halting to daily delivery so they can get the race going.

We did a large AI Readiness Assessment across dozens of PE portfolio companies this year.

Some were in a good place.  Some, not so good.

Almost universally, people did not understand Dark Launching. This is a critical capability for being able to deploy lots of code written with AI.

It was a wild year. I did a large AI Readiness Assessments on dozens of portfolio companies for a private equity firm along several different axes. Pa

Open article: Many are in the dark about Dark Launching (but it's an essential part of your AI toolkit) by Dave Mangot • 6 min read

Many are in the dark about Dark Launching (but it's an essential part of your AI toolkit)

If Patrick Debois is the godfather of DevOps, Mary Parker Follett might be the godmother.

In this lecture from January 1933 she talks about:

- Theory Y
- Intrinsic motivation
- Blameless postmortems 😮 
- and the importance of understanding how your work helps to achieve business goals and beyond.

Spooky!  Happy Halloween, Happy Friday.

It's been about 2 years since I published DevOps Patterns for Private Equity. With AI, the core problem I wrote about has only gotten worse.  For 2 weeks, let's dive into some of the battle-tested patterns I wrote about in the book.
-----
Day 1:  The Iceberg of Ignorance

Due diligence for deal teams is a "rock and a hard place" problem.  One the one hand, you want to collect as much high-fidelity information as possible.  On the other, you have to do it without everyone knowing about it.

Interviews hardly ever go below the leadership level meaning *we don't go to where the information is.* 

Cultural dysfunction is dark debt and it's a blind spot:

- who is an ITIL zealot in a cloud world
- whether EBS snapshots are deleted by governance automation or the quarterly "search of nickels everywhere" exercise
- that there is an executive who tells people they "aren't a team player" when they give him accurate information he doesn't want to hear

These are the things that make every minor improvement toward the investment thesis a major struggle.

We can't do much to change the due diligence process and that's ok.  The goal isn't to call off the deal, it's to fix these fixable problems.

Once we find them and recognize them.

That's your mission if you choose to accept it.  

(play Mission Impossible music)

#privateequity

I attended a really interesting webinar today hosted by the Association for Corporate Growth (ACG) featuring Pamela Hendrickson from the The Riverside Company.

She covered some things the many private equity folks are familiar with, and some they may not.

- if you focus on operational expertise, you will get better value creation

Elite operational expertise is what I work with PE portco CTOs on every day and it is the fundamental thesis of BayPine. Any PE professional who doesn't understand this yet hasn't been listening.

When the topic turned to AI, she said something I would not have expected.  She said if you want to be successful with AI, you should focus on:

- technology
- culture
- clean data

I highly doubt if she's read the 2025 AI DORA Report but that's spot on to what it says is essential for software companies to get real results from AI:

- Clear and communicated AI stance (culture)
- Strong version control practices (culture, technology)
- Working in small batches (culture)
- User-centric focus (culture)
- Quality internal platforms (technology)
- AI-accessible internal data (technology)
- Healthy data ecosystem (clean data)

😮 

I talk about these topics with engineers and CTOs all the time but did not expect to hear such an accurate summation in a one hour webinar for PE professionals.

Do you have those pieces in place?  Why or why not?

It was impossible for our industry to develop DevOps in the 90s.

Many of us have read Deming, Follett, and studied the theory of constraints.  Lean manufacturing had existed for decades. 

Why didn't the DevOps movement develop earlier?

If you're of a certain vintage, you remember people shifting from mainframes (you log into the machine "remotely") to client/server.

Eventually companies like Salesforce moved the server to "the cloud" (really data centers) which was the beginning of the end of on-prem software.

In the early days of the DevOps movement we used to talk about the problems of "throwing software over the wall" from Dev to Ops. (if you still have this problem, call me now)

Yes because:
 - the ops people didn't write it
 - the ops people can't fix it
 - it leads to more outages
 - etc

But underlying all that is Conway's Law.

With SaaS, operations is a part of the offering. Yes, we have code, but it exists WITH operations.  In this case, the communication patterns Conway talks about are the communication patterns between Dev and Ops. 

Separate them under different leaders and you will have a very bad time.  It's Conway's Law.

People started to recognize this in the early 2000s, but by the late 2000s, cloud was born and it became *painfully obvious*.

The late 2000s was when the DevOps movement was born.

Coincidence?  I don't think so.

The latest DORA report emphasized the need for good platform engineering if you want to be successful with AI. (if you need help with this, call me)

Coincidence?  I don't think so.

The way we develop and ship software is a part of the product.

Admit it.  Some of the funniest stuff you've read about Vibe Coding in the past year are people who thought that all you had to do was code something, and then they put it into production. Everything blew up.

You can't put software (especially enterprise SaaS) on the Internet without operations being a fully integrated part of that software.

Or best of luck beating Conway's Law.

Mary Parker Follet on "power".

Follet's thoughts on power were ahead of her time, as usual.

They were collaborative, making the distinction between power-over vs. power-with.

In order to really succeed in power-with, we need to move appropriate decision-making, authority, and responsibility to where the information is.   

Yes, you saw this idea in "Turn the Ship Around" almost 100 years later!  😮 

Happy Friday!

I worked with one portco that, through 20 acquisitions, inherited:

- 7 different database engines (Postgres, SQL Server, DB2, Oracle, Mongo...)
- 5 different monitoring systems
- 7 active data centers

This sprawl is a massive, invisible drag on what is supposed to be an internal platform if you want to get the benefits of your AI investment.

You're paying for redundant vendors, fragmented data, and crippling complexity.

You can't scale a museum.

There is little point in collecting all the technologies that your acquisitions have made in years past. 

You've collected the customers base, but also a product you don't know how to run and a bunch of old tech.

Instead of spending more money on DBAs, you need to spend that money on what you actually want to support.

How many database engines are you currently supporting?

#privateequity

taplio

Dave Mangot's Linkedin Analytics

Dave Mangot

Performance ScoreDave Mangot's LinkedIn Performance

Dave Mangot's Best Posts (last 30 days)

Ready to grow your LinkedIn presence?

taplio

.css-6ytdur{background-image:linear-gradient(284deg, rgba(39, 43, 72, 1) 0%, rgba(33, 47, 183, 1) 100%);color:transparent;-webkit-background-clip:text;background-clip:text;}Dave Mangot's Linkedin Analytics

Dave Mangot

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Dave Mangot's LinkedIn Performance

Dave Mangot's Best Posts .css-1emoo84{font-size:var(--chakra-fontSizes-md);font-weight:400;color:var(--chakra-colors-gray-600);}(last 30 days)

Ready to grow your LinkedIn presence?

Dave Mangot's Linkedin Analytics

Performance ScoreDave Mangot's LinkedIn Performance

Dave Mangot's Best Posts (last 30 days)