54 / 108 · 03 Agentic Testing Architectures · Case Study: The OpenObserve Council of Sub-Agents← prev⊞ allnext →☰ Read as one page
7.2Architecture: Eight Specialized Agents
+------------------------+
| COUNCIL ORCHESTRATOR |
| (Test Plan Director) |
+-----------+------------+
|
+-------+-------+-------+---+---+-------+-------+-------+
| | | | | | | |
[1] [2] [3] [4] [5] [6] [7] [8]
Code Test Test Test Flaky Cov. Doc PR
Anal. Gen. Run. Fix. Detect Anal. Gen. Rev.
| Agent | Role | Key Capability |
|---|---|---|
| Code Analyzer | Reads source code, identifies testable functions | Maps function signatures, dependencies, error paths |
| Test Generator | Writes new test cases | Uses code analysis + specification to produce Rust tests |
| Test Runner | Executes tests, captures results | Runs cargo test, parses stdout/stderr, reports pass/fail |
| Test Fixer | Repairs failing tests | Reads error messages, identifies root cause, applies fix |
| Flaky Detector | Identifies non-deterministic tests | Runs each test 5x, flags inconsistent results |
| Coverage Analyzer | Measures code coverage, identifies gaps | Runs cargo tarpaulin, reports uncovered lines |
| Documentation Generator | Creates test documentation | Produces test plan documents from the test suite |
| PR Reviewer | Reviews test PRs before merge | Checks quality, coverage, style conformance |