Robot Framework — Execution & Reliability
Short, practical notes on running Robot Framework faster (parallelism), recovering from transient failures (retries and native error handling), and keeping suites maintainable (setup/teardown and structure). Pair this with the official parallel execution guide and your CI documentation.
Topics in this section
| Topic | What you will learn |
|---|---|
| Parallel Execution with Pabot | Suite vs test-level splits, key pabot flags, isolation rules, PabotLib locks, tagging for selective runs, CI-oriented splitting. |
| Error Handling & Retry Patterns | TRY/EXCEPT (RF 5+), built-in recovery keywords, Wait Until Keyword Succeeds, fail-fast options, Python-side retries, failure logging. |
| Setup, Teardown & Test Organization | Suite/test/keyword setup scopes, guaranteed teardown behavior, realistic examples, tags and naming, robot CLI outputs and variables. |
Quick orientation
- Parallel runs buy wall-clock time only when tests are isolated: no shared files, DB rows, or browser sessions unless coordinated (for example with PabotLib).
- Retries belong around bounded, idempotent operations (API polling, element appearance), not around whole tests that hide product bugs.
- Teardown is where you release resources; setup is where you prove prerequisites so failures are obvious early.