janhoy opened a new issue, #129:
URL: https://github.com/apache/solr-orbit/issues/129

   launchr should be able to provision AWS machines for a benchmark run and 
store what it produces in S3, so the results outlive the instances. Two new 
capabilities, which together are what a scheduled nightly needs.
   
   **S3 as a storage backend.** The same storage contract as the local 
filesystem one — identical layout, identical code paths — so that switching is 
a single line of config, a laptop run can publish into the shared bucket, and 
reportr can build a trend site from results it did not produce. Writes should 
be immutable, so two sessions cannot silently overwrite one key, and a re-run 
gets a new id rather than clobbering its predecessor.
   
   **EC2 provisioning for a run.** A session should create its own instances, 
deploy Solr across them, run the load driver on a machine of its own, and 
destroy everything afterwards. Running the driver separately from the system 
under test is the point: it is what makes a number publishable rather than a 
measurement of whichever laptop produced it. Because the resources cost money 
for as long as they exist, provisioning needs a record durable enough to 
survive the process that created them, so a crashed session can be reclaimed 
later rather than discovered on a bill.
   
   Both pieces should go in behind the existing plugin registries — a storage 
backend and a provisioning strategy — so that Azure Blob Storage and Azure VMs, 
or GCS and GCE, are later additions beside them rather than rewrites of them.
   
   S3 comes first and stands on its own: infrastructure that is destroyed at 
the end of a session plainly cannot also be where the results live. It is also 
what #99 is waiting on before reportr can run anywhere but the machine that 
produced the data.
   
   🤖 Assisted by Claude Code


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to