Optimize the Disperser perf

[jianoaix]

Why are these changes needed?

The Disperser in particular the retrieval latency was 60+ seconds, when there was high traffic and/or large blobs.

There was due to inefficient memory management when dealing with blob buffer from S3, making it spend a lot of CPU/time on memory allocation and GC.

Tested:
From local retrieval query for a large blob,

Before:

time grpcurl -max-msg-sz 15000000 -proto ./api/proto/disperser/disperser.proto -import-path ./api/proto -d '{"batch_header_hash": "G9ntJR15dlOlljWQnx2avoT39XqLD5STAo91B8OMBqw=", "blob_index": 613 }' disperser-holesky.eigenda.xyz:443 disperser.Disperser/RetrieveBlob > /dev/null 2>&1
real	0m7.335s
user	0m0.357s
sys	0m0.157s

After:

time grpcurl -max-msg-sz 15000000 -proto ./api/proto/disperser/disperser.proto -import-path ./api/proto -d '{"batch_header_hash": "G9ntJR15dlOlljWQnx2avoT39XqLD5STAo91B8OMBqw=", "blob_index": 613 }' disperser-holesky.eigenda.xyz:443 disperser.Disperser/RetrieveBlob > /dev/null 2>&1

real	0m0.847s
user	0m0.382s
sys	0m0.152s

From pprof:

Before:

After:

Checks

• I've made sure the tests are passing. Note that there might be a few flaky tests, in that case, please comment that they are not relevant.
• I've checked the new test coverage and the coverage percentage didn't drop.
• Testing Strategy
  • Unit tests
  • Integration tests
  • This PR is not tested :(

[dmanc]

[ian-shim]

We already have this method in this file HeadObject. Could we have used that one?