How to decide what is worth building now that apps are cheap to create, and how to choose between apps, packs and bespoke configuration
1. You can build anything. Should you? Cribl apps
WATCH IN THE CRIBLCON 2026 KEYNOTE
44:06 The unseen work: deploy, govern, secure
1:01:36 Build vs buy
56:15 Marketplace launch
At CriblCon 2026, Cribl’s chief product officer typed a paragraph describing an application he wanted, pressed start, and went on talking. A few minutes later a working app was on screen, running on live data. He called it speaking apps into existence.
It was a good demo and the capability is real. It also changes the oldest question in IT, build or buy, because the cost of building has just fallen through the floor. The cost of owning has not.
Figure 1 – This figure shows four questions to ask before building an app; only an app that passes all four should be built without conditions.
1.1 What became cheap
Writing the first version. Cribl’s app framework gives an AI coding assistant a starter project, documented interfaces and a live preview, so someone who understands the problem can produce a usable tool without being a software engineer. Cribl says more than 150 apps have been built this way by its own teams, partners and customers.
For platform teams this is welcome. The request queue of “can we have a view that shows X for team Y” has always been longer than the team could serve.
1.2 What did not become cheap
Cribl’s product chief was direct about this. AI has made creating software far easier, but deploying, managing, securing and sharing it remain hard. The platform takes on some of that: hosting, access control and packaging are handled for you.
It does not take on the rest. Somebody still has to decide what the app is allowed to reach, review it before it is installed for other teams, fix it when the underlying data changes shape, and retire it when its author moves on. Fifty apps built in an afternoon each are fifty things with an owner, or fifty things without one.
Anyone who lived through the era of business-critical spreadsheets and unsupported Access databases will recognise the pattern. The tooling is better this time. The governance question is the same.
1.3 The line worth remembering
Cribl’s chief customer officer gave the best one-sentence answer of the keynote: having the power to build anything does not equal the responsibility to build everything. Build the things that are unique to your business. For problems everyone shares, start with something already built.
That is why the Cribl Marketplace matters as much as the build tooling. It is where apps from Cribl, partners and the community can be found and installed, with a review and certification process and a named party responsible for support.
1.4 A simple test
Before building, ask four questions.
- Is this unique to us? If another organisation has the same problem, look for an existing app first.
- Will it still matter in a year? If not, build it quickly and plan to delete it.
- Who owns it after the author leaves? No name, no build.
- What can it reach? If the answer includes sensitive data or external systems, it needs a proper review regardless of how easy it was to write.
1.5 What we would do
Set a light policy before the apps arrive, not after: who may build, who approves installation, where the inventory lives, and how often it is reviewed. Then let people build. The speed is a real advantage, and the organisations that benefit most will be the ones that made it safe to use.
Apto Feature: apto Operate covers app inventory, review and lifecycle as part of the managed service.
2. Apps, packs or bespoke: how to choose the right way to extend Cribl
WATCH IN THE CRIBLCON 2026 KEYNOTE
44:06 The unseen work: deploy, govern, secure
44:38 Customise, build workflows, reach beyond Cribl
Cribl Apps product page
There are now three ways to make Cribl do something it does not do out of the box. Choosing the wrong one costs either months of unnecessary engineering or years of maintaining something fragile. This is a short guide to choosing well.
Figure 2 – This figure shows the three routes for extending Cribl; effort to create rises from left to right while effort to reuse falls.
2.1 The three routes
Packs have been part of Cribl for years. Apps are new: Cribl’s framework lets you build an interface, workflow and visualisations on top of the platform’s data, search and access controls, and distribute the result through the Cribl Marketplace.
2.2 An example that shows the difference
Take a common request: “show me what is in our audit logs”. Done by hand, that needs a credential, a collector, a pipeline, a storage destination and a dashboard. Five separate pieces, each configured by someone who knows how.
A pack would deliver the pipeline and dashboard ready-made, leaving the credential and collector to set up and some tuning to do.
An app can collapse the whole sequence into one guided screen: enter the credential, press go, see the result. Cribl uses this example itself, and it is a fair illustration of where apps earn their keep.
2.3 When each is right
Stay bespoke when the need is a true one-off, nothing existing fits, and you do not expect to repeat it. Most early-stage work belongs here. Do not package something you have only done once.
Use a pack when a configurable template gets you most of the way, some local tuning is expected anyway, and a full interface would be over-engineering. Packs are easier to inspect and adjust, which matters to engineers who want to see exactly what is happening to their data.
Build or install an app when the workflow is repeatable, the people using it are not Cribl specialists, and a guided experience or custom visual would save them real time. Cribl’s product chief described the goal as software shaped around how your team works, not the other way round.
2.4 Three mistakes to avoid
Building an app because you can. The tooling makes it quick to produce something impressive. The useful test is whether it will still be used, and still be maintained, in a year. A few well-chosen apps are worth more than a long list nobody owns.
Forgetting the unseen work. Cribl is clear that the platform handles hosting, access control and packaging, and equally clear that deployment, governance and lifecycle still need managing. An app that reaches external systems needs a review before it is installed widely.
Hiding logic people need to see. A one-click app is ideal for an operations user and frustrating for an engineer debugging a parsing problem. For data-shaping logic that will be tuned often, a pack is usually the better home, even if an app sits on top.
2.5 A rule of thumb
Do it by hand the first time. Make it a pack the third time. Make it an app when the people asking for it are no longer the people who know how it works.
-
12 November 2026
The threat intelligence bill nobody budgets for
-
4 November 2026
AI Agents Just Exposed a Gap in Access Control; Here’s How the Industry Is Responding.
-
26 October 2026
Apps on Cribl: whether to build, and what to build
See how we can build your digital capability,
call us on +44(0)845 226 3351 or send us an email…





