I missed blogging yesterday as I was on stage/backstage for quite a bit of the keynote. Live updates, so keep refreshing.

Today I’m blogging and noting the interesting things I see. Today is the Redgate keynote, and disclosure, I work for Redgate, so take this with a grain of salt.

The opening this morning is from my good friend, Annabel Bradford. Annabel and I have been at events all over the world for over a decade, both Redgate hosted and third-party events, including many PASS Summits.

She’s welcoming everyone, sharing some memories from her previous 14 events, and giving a little housekeeping. It’s good to have a few reminders as there is so much happening. The

Annabel also reminds us to give feedback. Fill our evals and surveys, take pictures and post notes, and even email the Summit team if you want to see changes. As someone that has been on the back side of these events, we review and debate lots of feedback.

The Redgate Keynote: Simplifying Complexity: making the database work in the real world

The first speaker is Kellyn Gorman. She’s been a friend for around 15 years, formerly living in Denver. She’s come to SQL Saturdays and I’ve spoken at Oracle events, so we’ve crossed paths regularly.

The data world is scary, with so many data breaches. Lots of headlines.

And this is more painful because of multi-platforms. Lots of challenges above security, we have other issues, including just running systems

One of the possible solutions here is synthetic data, which shouldn’t be a security risk for data breaches. However, synth is hard, and Kellyn is leading to a new area where Redgate is spending some time.

It takes a village to get things running smoothly. She notes that AI and ML are here to help us, to be our assistants. She likes ML as a subset of AI. This will help us get more done in many ways, but one is with synth data.

Graham McMillan, CTO of Redgate

Graham is our new CTO, I met him this past summer. We had dinner in NYC a few months back, and I’ve gotten to see how he thinks as there is a weekly CTO report that he post.

He used to run an SaaS operation with 5,000 SQL Server databases. They were upgraded semi-rarely, and the environment felt brittle. He has been in the situation many of us have, trying to get things done, testing our systems, upgrading, and feeling the pressure from the business. He believes in investing in people, process, and tooling.

He introduces an IDC, an analyst from Europe. Archana Venkatraman. I find analysts interesting in that they do hear from lots of different organizations and people, but they can still be led astray by a few loud voices.

A few stats from what IDC sees from executives

• digital tech spend expanding 7x faster than the economy in 2024
• CEO expect 49% of rev to come from digital products
• 8-10 expect more than 25% of new apps to be cloud native

Excellence comes from speed, quality, efficiency. We know that, but we also know that lots of organizations aren’t doing well in these areas.

Lots of tech debt, especially after the pandemic where companies pivoted quickly.

Graham doesn’t like tech debt as a term as those were the decisions we made. I think that’s a bit short sighted, as we sometimes make conscious decisions we know are wrong, but that need to be made to get quickly.

3.5 hours on average to deploy to databases. That’s crazy.

I also believe it.

Why, lots of silo’d development, especially between DBAs/developers. The DBA world is often working in less-automated ways, often from historical habits.

Speed is a bottleneck, which lowers our productivity. Less time is spent on value activities, and more on firefighting, That’s something I see often, and something many DBAs don’t appreciate.

The Messy Middle

Made some changes to address low hanging fruit, but then get stuck. This is the messy middle from IDC. The image shows this in two ways, with infrastructure shown, as well as operational challenges.

Costs is a big one that is pointed out, especially the lack of awareness and controlling costs. This is a big deal in the cloud. Visibility is another.

Graham says we’re here as we make a decision, but we don’t follow through. That is truly a leadership and culture failure, not a tech worker failure. Archana repeats this, culture is a problem.

How can we do that? That’s the trick. Redgate has some ideas, and we’ll see how that might be done.

Arneh Eskandari and Danny de Haan, come on. Arneh is a solutions engineer I’ve worked with for well over a decade. He and I have debated, argued, informed, and presented often in the past about DevOps and how to get better at working with databases. We often agree and I enjoy our talks. These days, we are trying to build better engineers to demo and support customers.

He’s her with Danny, who helps run the databases for the largest pension company in the Netherlands. Their approach to making changes is evolved. Five years ago, they worked very manually. Each team had to request all their own changes, knowing what they needed.

They have moved to building a platform to make things easier. Some are bespoke, built in-house, but there are also third party products. That’s likely the reality for a strong org, build some things, buy some things. Things can be done in a portal, or through an API, to request resources (servers, services, cloud, local, etc.).

This makes things better in these ways.

• easy  (portal or API/code)
• independence (self-services)
• efficient, few people involved (self-service again)
• compliant by design (regulatory risk)
• secure by design (certain things allowed or disallowed)

Thoughtful automation – Arneh uses this term, which is what we should do. Don’t automate everything, and don’t just automate it for what the customer wants (or what the tech wants). Do it in a way that is compliant and secure by design, and give some choice, meaning API-first with a portal on top of this.

They have a baseline. They have an expected set of things that need to take place for infrastructure and code, meaning some limits/guidelines/restrictions. They flag things that deviate from the baseline. That’s an approach that I like, since there are reasons to sometimes go away from a standard.

However.

Most of the time the deviations are simple mistakes. We’re too far along in tech to allow simple mistakes to slip through often, which is likely in the era of constantly changing staffs. Use automation to flag these, but have a way to create exceptions if needed.

More Industry Insights

Graham and Kellyn come back.

“Devops is essential. If you’re not automating, you’re not doing a good job.” – Kellyn.

Graham notes we run the State of Database Landscape survey. We’re run this quite a few years, and this year (the 2025 report from 2024 research), we had 2500+ responses. Report coming in Jan, but a few early results.

What do we want. What’s efficient? Kellyn says update. She’s prefer a db with no users, but that’s not possible (laughs).

She wants to patch or upgrade the database platform, but Kellyn doesn’t want to do changes for deployments. She does like tuning.

74% using more than one platform, along with organizations trying to incorporate lots of new technology. 26% using more than 4 platforms. Audience survey, but I couldn’t see how many responded.

18% (almost 1 in 5) making daily changes. There is also a 50% increase in short term changes between 2022 and 2024. These are bug fixes, security fixes, or enhancements. I see pressure to do this, but I also worry about this. Short term can be rash changes if not thought through.

The mission of Redgate is to help here. We display this.

The

Redgate bought db-engines, which Kellyn has used for years. She shows growth across some top platforms, which is similar to what we showed last year. We’ve highlighted this at a few events this year.

Flyway supports databricks, Clickhouse and cassandra. MongoDb as well. N

Not just a breadth though, but we are adding depth. Lots of changes in Flyway, as well as Redgate Monitor, which now runs on Linux/Timescale (PostgreSQL) if you want.

We’re into the product listing.

TDM mentioned, which was announced last year, and they note we’re improved subsetting performance. I am happy to see this (I knew this as I’ve been helping the team lightly), because I think this is crucial for success.

Cloud Movement

Kellyn is showing the the all in the cloud numbers have gone down in 2024 from 2023. Why? There is an increase in all on-premises environments. This matches Kellyn’s experience as she has worked on a lot of VLDBs, which can’t work in the cloud. She’s seen some of the large systems struggle to work in the cloud and they’ve come back.

That’s not most of us, but it is some of us. I never thought VLDBs made sense for the cloud if they were heavy on compute and IO. If it’s just size and they can tolerate latency, great.

Initial investment in cloud movement is low, but if you don’t manage this, cost increases can get out of control. Another reason some orgs might be moving back.

Cloud marketplaces are the place for Redgate. We’ve added Flyway, Monitor and TDM in the cloud.

Biggest challenge from the audience? Data integration issues. The report, skillset requirements and training. However, this is an audience whose companies invested in them for the most part, so I think skills are less an issue here. They have more tactical challenges.

Back to products, with James Hemson. He helps run TDM and I talk with him weekly. The first item to look at is security. That has limited 21% of organizations to adopting other platforms. I guess that makes sense, though I don’t know new platforms are needed. Maybe.

However, that’s a business decision. We don’t want security to be the driving force unless the new db platform is insecure. In that case, don’t use it.

That brings us to Redgate Monitor Enterprise. We launched this earlier this year as a new tier. One of the things added in Ent version was compliance checking, where we look at configuration against a baseline. We also added some security information on access to servers and databases. James shows some of this.

I won’t, but you can see this online for permissions and compliance in our demo system. A nice quote: you can see you’re compliant all the time, not just at audit time.

People believe in AI. It improves productivity (at least a belief in our survey). Hmmm, not sure I think this, but I’m skeptical.

AI can help you in some ways. Redgate has been working to add AI to Monitor. Example, lots of alerts. Too many. If we lowered a threshold, we can remove a lot of noise. We can get alerts when things are very or unusually high, not just higher.

ML is being used to tailor alerts to limit them to when they’re important or needed, not just every alert. This is a great use, since many of us don’t need to keep tuning alerts. We need to respond to them.

I wasn’t aware of what Monitor was doing here, but I like it.

We have a lack of collaboration. Bringing Dev and Ops together, a la DevOps, should have solved this.

A bit of a stretch to me, but I was aware of this. We wanted to ensure we can get test data to devs easier, without a lot of Ops effort. We are adding beta synth data into TDM as of today.

I’ve seen this, and it’s got possibilities, but I’m still unsure of how effective this will be. You judge, but there are some interesting ideas with ML looking at your data and generating synth data that looks like your data, but it’s not real. Great idea, but we’ll see how well it works for lots of different data structures.

There are both rules-based and AI-based versions of this that can be used in different ways. It was an interesting idea when I saw it a few months ago, but I’ll leave it there. You test it and let me know if it works for you.

Monitor, automate, protect, the three areas where Redgate works. Depending on your job and your challenges, you might use one of these.

Was this keynote that showed how to simplify complexity in the real world? I think there were some good things here, though there are certainly some product focuses here on Redgate solutions, but the high level concepts are things that help us make the world better.

Not simple for all the people doing the work, especially in Operations, but for the users, life should be simple. Watch the replay

Final announcements, not just 1 Summit next November, but 4. Three more in other cities (NY, Chicago, Amsterdam), which will be smaller and shorter (1-2 days). I’ve known about these, but wonder if you like this.

Come to the Redgate booth and win an Aardvark, which represents the first product from Redgate. If you get the golden one, you get $250 as well.