1.8.1

v1.8.1

Make ami_id configurable (#357)

1.7.1

Update postrun.json when job fails (#353)

* Update postrun.json for failed runs

* Update postrun on idle termination

* Remove unnecessary datetime

1.7.0

Add possibility to use Caper to execute WDL files.

1.6.0

Improved spot failure detection. Added retry logic when pulling the docker image.

1.5.0

• ECR support for tibanna awsf docker image

1.4.1

• Now we have options for deploying lambdas into specific subnet / security group.
• Singularity version is updated to 3.8.1-ce.
• Log includes the contents of the workflow run directory

1.3.1

• tibanna kill now has a soft-kill option -z which does not immediately abort the execution but sends an abort signal to be picked up by the check_task lambda (the instance is terminated immediately) - this way the execution will have a FAILED status with a JobAborted exception (as opposed to a hard kill which immediately stops the step function execution which gives it an ABORTED status). This options could be useful for a wrapper (e.g. tibanna_ff) to be able to handle the abort more gracefully.
• A newly introduced issue of execution ARN dependencies causing error is now fixed.
• tibanna cost_estimate now has a force option -f which estimates the cost even if the actual cost is available.

1.2.8

• cost_updater lambda name prefix no longer hard-coded for better extensibility (no change in behavior)

1.2.7

• jobid can be provided in input json (can be overwritten by run_workflow's jobid option)
• run_workflow response's response->executionArn points to the correct execution not the costupdater execution.

1.2.6

• Deployment now does not fail when deletion of S3 bucket public access block failed - just giving a warning.