Explore your software
Knowledge shows up in two places in the console. The Knowledge Base section gives you the whole organization: search, impact, patterns, and open questions. Each application also gets its own Knowledge Base view with its book, relations, and facts.
Everything you read links back to its evidence. A relation shows the call it's based on, a fact shows its source and commit, and a book chapter links to the lines of code it describes.
The Knowledge Base section
Open Knowledge Base from the switcher at the top of the console, next to Development. The menu on the left holds every page of the section.
Search and ask
Use the search box on the Overview or on Explore. Results are grouped by kind, such as documents, components, facts, relations, questions, and knowledge, and you can filter by kind with the chips above the list. Select a result to read it next to the list, then jump to its application or to its impact from the reader.
When you type a question, such as "how do we send a transactional notification to a customer?", an AI answer may appear above the results. It reads the most relevant results and answers by citing them, so you can open each source. It never writes anything to Knowledge.
AI answers can be wrong. Check the cited sources before you act on one. If Knowledge doesn't hold the answer yet, it says so and shows you the closest results instead.
With an empty search, Explore lets you browse two things:
- Norms and know-how: everything your organization wrote down, grouped by where it applies, from Everywhere down to For one application. This is also where you add your own know-how.
- Facts by application: the structured facts the analysis wrote about each application, readable as tables or YAML, with the source and commit of each one.
Check what a change reaches
Impact answers "what breaks if this changes?". Pick a shared component, such as an internal API, a library, a vendor, a queue, or a datastore, and see everything that depends on it, grouped by namespace.
The panel on the right ranks the calls that reach it: the HTTP method and path each consumer uses, and how many components make each call. A change to the call at the top reaches the most consumers. Select a group or a dot in the map to list its components, and See its impact on any of them to follow the chain one more hop.
Use Impact before you change a contract, rotate a shared credential, or upgrade a core system. It turns "who do we have to warn?" into a list with the operation each team uses.
Fix a problem once with patterns
Patterns groups the findings that repeat across applications, such as the same misconfiguration copied into a dozen services. The chart places each pattern by how far it reaches and how severe it is, and Start here: wide and severe is the corner to tackle first.
A pattern's page shows where it comes from, for example the same lines of code copied from one source, how it's fixed, and every application that has it. Click an application to open its action items already filtered to that pattern.
Browse the applications
Applications lists every analyzed application with its criticality, its open findings, its open questions, how many components use it, and how old its analyzed code is. Use the quick filters to find the ones with high-severity findings, open questions, or code older than six months.
Learning path
When your organization has it, Learning path builds a reading list for you. If you're new, pick a goal, such as Onboarding, Ship changes to my apps, or Be on call, and the namespaces you'll work in. You can also ask for a path about any topic, such as "how approvals work". It picks the book chapters, runbooks, norms, and open questions that matter for that goal, and tracks what you've read.
If you've been around for a while, it also shows the applications you changed recently and the people who changed them with you, so you know who to ask.
An application's knowledge
On an application's overview, the Knowledge Base card on the right sums up what Knowledge knows: its criticality, how old the analyzed code is, its open findings and questions, and who uses it. Click Open, or open Knowledge Base from the application sidebar, to read the full view.
The view has five tabs:
| Tab | What you'll find |
|---|---|
| Overview | The book, chapter by chapter, and the other documents, such as the runbook and its risks. Search or ask about this application from the box at the top |
| Relations | What the application calls and what calls it, as a map. Choose a lens: Runtime (APIs, queues, identity providers, datastores), Deployment (the scopes it runs on), Build & supply chain (base images, libraries, CI templates), or All. Turn groups on and off under Depends on and Used by to focus the map. Select a component to see exactly which endpoints it calls |
| Questions | What its analysis couldn't answer from the code. Answer them from here |
| Profile | A summary of its facts: where it runs, what it exposes, how it's configured, the data it holds, access and secrets, and the standards it fails |
| Facets | Every fact recorded about it, with its provenance and when it was updated. Search them or filter by provenance, and open one to read it as YAML or as tables |
In Profile, open What it exposes to see every endpoint with the permission it checks, whether it's public, and the file where it's defined.
If an application has no knowledge yet, it hasn't been analyzed. Its knowledge appears after the first analysis.
Dependencies on the application overview
Your administrator can also add a Dependencies card to every application's overview. It shows the application's relations as a small map and a table, and See all relations opens the Relations tab.
Related docs
- Contribute knowledge: answer open questions, add know-how, and correct what's wrong
- Connect your agents: ask the same questions from your coding agent
- Triage and resolve action items: work through the findings Knowledge files