BFD-3664: Pipeline job for SAMHSA tag backfill. #2506
Conversation
…if/samhsa/CcwTagKey.java Co-authored-by: aschey-forpeople <[email protected]>
dondevun
requested review from
mjburling,
aschey-forpeople and
malessi
as code owners
| ConfigLoader config, boolean ccwPipelineEnabled) { | ||
| boolean enabled = config.booleanOption(SSM_PATH_SAMHSA_BACKFILL_ENABLED).orElse(false); | ||
| // We don't want to run if we're on a CCW Pipeline instance | ||
| if (!enabled || ccwPipelineEnabled) { |
There was a problem hiding this comment.
It is a bit confusing that this runs on the RDA pipeline, but I understand the reasoning. Ideally this could run as its own pipeline, but that would increase the complexity a fair bit. I think this is fine for now and we can revisit once we're running in ECS.
| String queryStr = | ||
| strSub.replace( | ||
| startingClaim.isPresent() ? QUERY_WITH_STARTING_CLAIM : QUERY_WITH_NO_STARTING_CLAIM); | ||
| return entityManager.createNativeQuery(queryStr, tableEntry.getClaimClass()); |
There was a problem hiding this comment.
Would it be possible to use createQuery instead of createNativeQuery? That will at least perform some type checking to ensure the type is assignable to the query result. I believe that would remove the need for select * as well.
There was a problem hiding this comment.
Yeah, we can do that. The check for existing tags may be a bit tricky, but I'll see what I can do.
There was a problem hiding this comment.
I think these changes should work fine. Testing performance now.
| @@ -152,6 +152,8 @@ | |||
| /bfd/${env}/pipeline/nonsensitive/rda/cleanup/run_size: UNDEFINED | |||
| /bfd/${env}/pipeline/nonsensitive/rda/cleanup/transaction_size: UNDEFINED | |||
| /bfd/${env}/pipeline/nonsensitive/rda/instance_type: m6a.large | |||
| /bfd/${env}/pipeline/nonsensitive/rda/samhsa/backfill/enabled: "false" | |||
| /bfd/${env}/pipeline/nonsensitive/rda/samhsa/backfill/batch_size: 15000 | |||
There was a problem hiding this comment.
Might want to test with a larger batch size here to see if it helps. 100,000 or so shouldn't be a problem.
|
|
dondevun
force-pushed
the
BFD-3664
branch
2 times, most recently
from
890887b to
b75b481
Compare
JIRA Ticket:
BFD-3664
What Does This PR Do?
This PR adds a pipeline job to backfill the SAMHSA tags tables. This will be able to run concurrently with the RDA pipeline job, but is disabled on CCW pipeline instances.
This will process all of the tables that could have SAMHSA codes concurrently, each with its own entityManager. There were some tradeoffs for the sake of performance, the biggest being that it does not construct entities from the SQL queries. Instead, it returns arrays of objects, and relies on the code to be aware of the types in each array position. This is obviously not ideal, the biggest issue being possible ClassCastExceptions if the array objects are not processed in the correct order; however, without the entity class to map the columns to the types, it is unfortunately unavoidable for this implementation.
What Should Reviewers Watch For?
If you're reviewing this PR, please check for these things in particular:
What Security Implications Does This PR Have?
Please indicate if this PR does any of the following:
Adds any new software dependencies
Modifies any security controls
Adds new transmission or storage of data
Any other changes that could possibly affect security?
I have considered the above security implications as it relates to this PR. (If one or more of the above apply, it cannot be merged without the ISSO or team security engineer's (
@sb-benohe) approval.)Validation
Have you fully verified and tested these changes? Is the acceptance criteria met? Please provide reproducible testing instructions, code snippets, or screenshots as applicable.