Challenge
Lobster is Germany’s pioneering data integration company with a clear purpose: connecting people and data for a better future.
The Lobster Data Platform unifies data integration, orchestration, and insights, to connect modern and legacy systems. It helps to turn information into smarter decisions. With its AI connectivity, organizations can enrich existing data, extend AI to legacy environments, and drive intelligent automation and analytics. Trusted by 2,000+ customers worldwide, Lobster streamlines operations and accelerates value from data.
However, internally, the process of provisioning and managing infrastructure was anything but simple. The documentation was inconsistent, automation unreliable, and essential knowledge lived in a single engineer’s head. As demand grew and with CentOS 7’s end-of-life approaching, Lobster’s cloud operations team needed to modernize—fast.
What began as a focused effort to streamline provisioning evolved into a company-wide transformation. With CloudBolt, Lobster moved to a standardized, blueprint-based model that reduced manual work by 90%, eliminated configuration drift, and enabled safe self-service across departments. “Our initial goal was just to install our software in the cloud,” said Ludwig Müller, Lobster’s cloud engineering team lead. “What we got in return from CloudBolt unlocked way more than we ever expected.”
“Our initial goal was just to install our software in the cloud. What we got in return from CloudBolt unlocked way more than we ever expected.”
Why CloudBolt
Lobster evaluated several platforms, but ran into trade-offs. Red Hat Ansible Automation Platform offered depth, but required a complex, multi-component deployment just to get started. MAAS.io was tightly tied to Ubuntu whereas Lobster’s systems ran on Red Hat Enterprise Linux. Cloudify, on the other hand, relied heavily on YAML, which made even basic scripting painful to do.
CloudBolt gave the team what they needed: a structured platform that worked with their existing tools and let them move at their own pace.
“CloudBolt was the sweet spot for our team because it was accessible yet powerful,” Ludwig said. “It let us take tools we already knew—Ansible, Terraform, Python—and connect them like Lego pieces. No rewrites, no ramp-up. Just plug in and go.”
That flexibility lowered the barrier to adoption which, according to Ludwig, is the most important requirement of all. “What we’ve learned throughout this whole process is adoption beats architecture,” said Ludwig. “You can have the best sounding architecture, but if no one uses it—it’s worth nothing.”
What we’ve learned throughout this whole process is adoption beats architecture. “You can have the best sounding architecture, but if no one uses it—it’s worth nothing.”
Outcomes
Provisioning: From multi-team effort to 15-minute blueprint
Lobster’s first priority was to streamline provisioning. The team anchored the process on the customer’s contract ID. That four-digit input triggered an internal API call, returning everything the system needed to build the environment: product size, support tier, backup retention, region, VPC configuration, and tagging conventions. CloudBolt blueprints exposed only a few options—test vs prod, monitoring on or off, backup duration—and everything else was automated.
What used to take 4-6 hours of effort across multiple teams now happens in 15-45 minutes.
With this system, what used to take 4–6 hours of effort across multiple teams now happens in 15–45 minutes, depending on customer size. “After 15 minutes, we get an email with the URL to the platform that the customer ordered, and the customer gets an email with a password to start using the platform after 15 to 20 minutes. No technician needed,” explained Ludwig. “The only thing our teams needed was the contract ID. CloudBolt handled the rest.”
Day-2 operations: 30K jobs/month, no ticket required
Once provisioning stabilized, the team turned to Day-2 operations. They used CloudBolt to define scoped actions for restarting services, pulling logs, updating licenses, installing agents, and triggering platform pushes. Support could now perform many of these tasks directly without filing tickets or waiting on engineering. “Before, even pulling logs meant going through the customer, the networking team, and a help desk ticket. Now Support just clicks a button,” said Ludwig.
The team also integrated CloudBolt with PagerDuty and their legacy ticketing system so Support can get visibility into environment-level alerts and history without leaving the interface. Today, CloudBolt runs ~30,000 human-triggered jobs per month with a 90% success rate. Most of those actions now happen outside of engineering.
Company-wide adoption: From CloudOps to every team
Meanwhile, usage expanded across the business. Sales Ops began provisioning demo environments directly, cutting lead time from two weeks to one day. The internal Academy launched training labs on demand. Developers reused production-grade blueprints for test systems. What started as a CloudOps initiative became the default infrastructure layer for the entire company. “At this point, there’s not a single department at Lobster that doesn’t touch CloudBolt,” said Ludwig. “It wasn’t something we pushed—people saw the results and wanted to use it themselves.”
“At this point, there’s not a single department at Lobster that doesn’t touch CloudBolt. It wasn’t something we pushed—people saw the results and wanted to use it themselves.”
Scalable growth: 5x expansion without the overhead
Since adopting CloudBolt, the CloudOps team has gone from three engineers managing 100 services and databases across the data center and cloud to five engineers overseeing 500-600 servers, 450 databases, and 40 AWS accounts. With just two additional people, they’ve scaled their operational footprint by more than 5x, without adding complexity or risk.
“Now two of us can go on vacation at the same time and everything is covered,” said Ludwig. “That used to be impossible.”
“At this point, there’s not a single department at Lobster that doesn’t touch CloudBolt.”
Results
~90% reduction in manual effort
6x reduction in provisioning time
500+ servers and 450+ databases across 40 AWS accounts
~30,000 jobs/month—mostly by non-engineers
Configuration drift nearly eliminated
Support now handles Day-2 tasks without tickets
Company-wide use across support, sales, engineering, and academy
Conclusion
Lobster didn’t set out to overhaul its operations. It just needed to fix provisioning.
But with CloudBolt, they ended up with a platform that eliminated bottlenecks, reduced dependency on tribal knowledge, and scaled seamlessly across the business. It worked for their engineers—and just as importantly, for everyone else.
See how I&O teams are automating delivery, enforcing policy, and scaling hybrid operations—without giving up control.
Request a demo