I understand the complete pipeline again. I'm asking the same question: Is there an optimized solution to avoid dropping a completion status file to an S3 folder, and instead trigger a second Glue function from the first one?

In Glue you can avoid writing a status file by using job chaining or triggers. One common pattern is to configure the first Glue job to emit a CloudWatch event or an EventBridge rule when it completes. The rule can target the second Glue job directly, eliminating the need for an S3 marker. Alternatively, you can use Glue job bookmarks to track processed data and let the second job run only on new data. If you need more control, wrap the two jobs in a Step Functions state machine: the first job runs, and on success the state machine invokes the second job. This approach gives you retry logic, error handling, and visibility. Each method keeps the pipeline event‑driven and removes the file‑based hand‑off, improving reliability and reducing latency. The choice depends on your orchestration needs: simple triggers for linear pipelines, or Step Functions for complex workflows.