[#2525][FOLLOWUP] fix(server): remove metric buffer_block_size

[xianjingfeng]

What changes were proposed in this pull request?

Remove metric buffer_block_size

Why are the changes needed?

The observations of Summary are expensive due to the streaming quantile calculation and synchronized is used in Summary.observe.

"epollEventLoopGroup-3-130" #341 prio=10 os_prio=0 cpu=51979.29ms elapsed=2103.78s tid=0x000004c0882e7000 nid=0x486d waiting for monitor entry  [0x00007f21769e7000]
   java.lang.Thread.State: BLOCKED (on object monitor)
        at io.prometheus.client.TimeWindowQuantiles.insert(TimeWindowQuantiles.java:36)
        - waiting to lock <0x00000003006a7df8> (a io.prometheus.client.TimeWindowQuantiles)
        at io.prometheus.client.Summary$Child.observe(Summary.java:283)
        at io.prometheus.client.Summary.observe(Summary.java:309)

Does this PR introduce any user-facing change?

No.

How was this patch tested?

CI

[github-actions]

Test Results

 3 082 files  ±0   3 082 suites  ±0   6h 49m 33s ⏱️ + 1m 49s
 1 193 tests ±0   1 192 ✅ ±0   1 💤 ±0  0 ❌ ±0 
15 123 runs  ±0  15 108 ✅ ±0  15 💤 ±0  0 ❌ ±0 

Results for commit b7391b7. ± Comparison against base commit 066e71e.

[zuston]

lgtm

[xianjingfeng]

@zuston Thanks for your review.