In case anyone is interested, this AWS Batch runni...
# infra-deployment
m
In case anyone is interested, this AWS Batch running the Meltano container is running along pretty seamlessly for the past 3 months (linked is the template) https://github.com/mattarderne/meltano-batch
Goal was simplest reliable Meltano ELT. Doesn't include that much logic around retries and alerting, so could be improved there. I heard from someone that they are running Meltano in GIthub actions, which would certainly be an improvement in the sense of simplicity
t
Nice! Thanks for sharing 🙂
f
@matt_arderne for your
aws batch list-jobs
you can add something like this:
--filters '[{"name":"JOB_NAME","values":["prod_*"]}]'
to list all jobs that start with prod_. Otherwise, it only shows RUNNING jobs. Thus you can get a list of all jobs where the job name starts with prod_. You can also use something like
--job-status "SUCCEEDED"
instead, which will return all jobs with a SUCCEEDED status, but you can only specify one job status in the call. So it depends what you want to list, SUCCEEDED jobs, or maybe FAILED jobs, use job-status; all jobs regardless of state, use --filters. You can also use --filters to filter things like BEFORE_CREATED_AT and AFTER_CREATED_AT, as well as JOB_DEFINITION. But you can only use one filter name/type at a time.
Without having to spin this up myself, what is the output of the lambda invoke call in the "Create a Job" section? Is that the output of the Lambda function itself, or the Meltano job? Help for the invoke function says that LogResult is "The last 4 KB of the execution log, which is base64 encoded." This makes me think it is the lambda function and not the meltano job.
Ah, the output of the meltano job should show up in the Batch output, which goes to CloudWatch Logs...