Add component description for merge-jobs

src/merge-jobs/README.md

@@ -0,0 +1,55 @@
+# merge-jobs
+
+The **merge-jobs** component is responsible for taking result messages from the queue,
+download the attached MBTiles file and merge it into the latest planet MBTiles file.
+
+![Flow Diagram](merge-jobs-flow-diagram.png)
+
+## Usage
+
+### Docker
+
+To use **merge-jobs** inside the Docker Compose workflow you only need to ensure
+that you have a `planet.mbtiles` ready in the `export` folder.
+
+Now **merge-jobs** will merge the MBTiles files into `./export/planet.mbtiles`.
+
+```
+docker-compose run merge-jobs
+```
+
+### Python
+
+You can execute **merge-jobs** CLI directly as well without the wrapper Bash script.
+You can also configure the name of the queue to pull messages from
+with `--result-queue=<result-queue>`. You need to set the MBTiles merge target explicitly
+via `--merge-target<mbtiles-file>` to the MBTiles you want the MBTiles job results merged into.
+
+
+```
+merge-jobs.py <rabbitmq_url>  --merge-target=<mbtiles-file> [--result-queue=<result-queue>]
+```
+
+## Troubleshooting
+
+### Check Queue Length
+
+To get started you will need to have result messages in the message queue.
+
+```bash
+# Check queue length on local RabbitMQ instance
+rabbitmqctl list_queues
+
+# Or attach to Docker container running RabbitMQ
+docker exec <rabbit-container-id> rabbitmqctl list_queues
+```
+
+### S3 Files not accessible
+
+The **merge-jobs** component does need **public access** to the files referenced
+in the results message. If the bucket is not public you need to explicitely
+make all object inside the bucket public.
+
+```bash
+s3cmd setacl s3://osm2vectortiles-jobs/ --acl-public --recursive
+```