Fridolín Pokorný

Let a flow fail even if fallback was run

3

It would be nice to let a flow fail even if there was a failure in the flow and a fallback was started: ```yaml failures: - nodes: - task1 fallback:...

Let's move cache definition from flow to nodes.yaml. This way a cache instance wouldn't be bound to the particular flow but rather can be reused across different flows.

As Selinonlib is now merged with Selinon, there is no need to translate configuration to Python code, write it to a file and after that load and parse it again....

For the following configuration: ```yaml tasks: - name: A import: foo.bar - name: B import: foo.bar - name: C import: foo.bar - name: D import: foo.bar flows: - flow1 flow-definitions:...

It would be nice to support task timeouts so task execution time can exceed: ```yaml tasks: - name: 'Task1' import: 'myapp.tasks' timeout: 600 ``` In that case a special exception...

It would be nice to add support for `store_error_only` configuration option. Using this option will ensure that actual results are not stored in a storage, but rather only errors are...

Explore possibilities to enhance logging of tracepoints using daiquiri [1]. This enhanced logging lib can be also used in Selinonlib. See [2] for examples. Thanks to @jd. [1] https://pypi.python.org/pypi/daiquiri [2]...

Inspect results of parent subflows in edge conditions

1

Currently, one of the major limitations Selinon has is the fact that results of parent sub-flows cannot be used in conditions. It would be nice to include such capability.

Let make possible to propagate finished and failed nodes from subflows. This way a user can get information about finished and failed nodes in parent subflow.

Log to stdout/stderr

1

We run postgres deployment from OpenShift service catalog on RH PSI. The postgres container logs to files which makes it harder/not straightforward to check what's going on in the OpenShift...