Increase memory and timeout configuration for the transform lambda

We started seeing out of memory errors in the transform lambda
when onboarding the variant-nowcast-hub. This is the first time
we've onboarded a hub to S3 that includes sample data, and it
looks like the default memory allocation of 128MB isn't
sufficient.

For example, a single UMass submission is 27MB in memory after
the parquet file is read by pyarrow. We can and should
investigate ways to make the transform process better about
memory, but let's increase the memory now to ensure we're
in a good place until then.

src/hubverse_infrastructure/shared/hubverse_transforms.py

@@ -169,7 +169,8 @@ def create_transform_lambda(
         s3_bucket=s3_bucket,
         s3_key=s3_key,
         tags={"hub": "hubverse"},
-        timeout=210,
+        memory_size=500,
+        timeout=600,
         opts=ResourceOptions(depends_on=[hubverse_asset_bucket]),
     )