Few minor diagnostics perf improvements

[FabianMeiswinkel]

Description

Few minor perf improvements with diagnostics enabled that were identified when going through CPU profiles.

No monitoring, no Java Agent

About 5 percent throughput improvement with the changes in this PR - primarily by avoiding to use ImmutableEnumMap (which has allocation overhead in iterators)

https://fabianm-perf-results.westeurope.kusto.windows.net/fabianm-java-perf-results?query=H4sIAAAAAAAEAG1RS2vCQBC%2bC%2f6HYU8JGI22QgmkoKUHS7USpdcySSbJFHdTdtdX6Y%2fvmkgV6p5mZ%2fZ7zLc5WkzRkCcKTBmVDD5xh8EX6SLQZLYba4TfT9qq2%2fmBfUWaYD2bP6%2fWk%2fkSHgHL2hvn%2fmX45tBouVYQxyASwvwVLansKG4T5G5qWZI3Ckf3QTgMwjEMH6K7MArDftieK%2fqpRpVVC5QExqK2Zs%2b2AlFrLllFW0PaDM7LXAuSse%2bouTHWgFmB17j7WNTzWrF1DKp0lxeXwKQkZUWjSgdnPofVtij4EBurM7Qec%2bFdjMTxTfmBRFbZplYkeiCmLuYNt%2fW6YhMsE%2bG7OnKNp1pKtrP8WrAhPsu14ufH%2f1ZpUGYrpWt%2bE%2bCudIAsI2MSF60P6RFSVt5f5j0YSqd8gp6Q2qmRhtMXZJULtNvpdn4BfO%2fL1hgCAAA%3d&web=0

database("fabianm-java-perf-results").Results
| where TIMESTAMP > ago(5d)
| where Operation == "ReadLatency"
| where TIMESTAMP > datetime(2024-01-05 18:30:00.0000000)
| where BranchName startswith "origin:users/fabianm"
| where TestVariationName in ("Read_NoMonitoring_NoJavaAgent")
| extend Suffix=strcat(iif(BranchName=="origin:users/fabianm/mainclone", "Baseline", "This-PR"), ":", CommitId)
| extend Name=strcat(Suffix, ":", TestVariationName)
| summarize avg(SuccessRate) by bin(TIMESTAMP, 1m), Name
| render timechart

Metrics only (tracing disabled) via Java-Agent (but with Reactor disabled)

Even the other micro-optimizations (avoiding string.format etc.) helped to increase throughput slightly. The throughput improvement was relatively small, but continuously better than the main-branch. With the change to avoid ImmutableEnumMap the improvement across the board increased.

https://fabianm-perf-results.westeurope.kusto.windows.net/fabianm-java-perf-results?query=H4sIAAAAAAAEAG1RS2vCQBC%2bC%2f6HYU8RjEZboQgpaOnBUh%2botMcybibJlOym7G58lP74rlGqUPc0OzPfY2YSdLhBS4FIccOoVfiJWwy%2fyKShIVsVzopWZ3mKmo0f2OVkCNaT6fNqPZou4BEwK4NB0roU5x6NjksNcQxiSZi8oiMtD%2bI2QeKrjhUF%2fah%2fH0a9MBpA72F4Fw2jqBOd3hX92KCW%2bQwVgXVonN2xy0GUhjPWw8qSsd3zMNeCZN0bGq6N1WDWENTuPt49wZScYWnnujjU%2fxe%2fhlFG2o0rNyt9m3SlEbUN2vtpElhVacr72Doj0QXMaXBxFsc3%2fXQVspZFqUm0QYz93gs%2bxeucbbhYipaPhz7xVCrFbpJcC9bEZ7mT%2bLn532w1ylZK%2beQ3AW4zD5CSrF36Xbdgc4AN6%2bDvCG3oKa98hB6RxquRgeNNZO433Gw0G79Bcg%2fVKQIAAA%3d%3d&web=0

database("fabianm-java-perf-results").Results
| where TIMESTAMP > ago(5d)
| where Operation == "ReadLatency"
| where TIMESTAMP > datetime(2024-01-05 18:30:00.0000000)
| where BranchName startswith "origin:users/fabianm"
| where TestVariationName in ("Read_WithMetricsOnly_WithJavaAgentButNoReactor")
| extend Suffix=strcat(iif(BranchName=="origin:users/fabianm/mainclone", "Baseline", "This-PR"), ":", CommitId)
| extend Name=strcat(Suffix, ":", TestVariationName)
| summarize avg(SuccessRate) by bin(TIMESTAMP, 1m), Name
| render timechart

Metrics only (tracing disabled) but No Java Agent (or other MeterRegistry subscriber - so, Global MicroMeterRegistry is actually empty)

Before this PR even when no actual meter registry existed, the SDK code to populate metrics was stille xecuted - the actual Micro meter APIs to register values are no-ops in this case, but the overhead was still visible.

https://fabianm-perf-results.westeurope.kusto.windows.net/fabianm-java-perf-results?query=H4sIAAAAAAAEAG1RS2vCQBC%2bC%2f6HYU8JGI22QgmkoKUHS30QpT3KJJkkU5JN2V1fpT%2b%2bayJVaPc0zMz3mG9TNBijJkdkGDPKyvvAPXqfpDJPkd6VRgu3H7VVt%2fMNh4IUwWY2f15vJvMVPALmtTNO3etwadFouJYQhiAiwvQVDcnkJP4nSO3UcEXOyB%2fde%2f7Q88cwfAju%2fMD3%2b377buinCmVSLLAi0AaV0Qc2BYhacc4y2GlSenA55laQtHlDxY2xBswSnMbd9t0SzMkoTvRSlqfton6xIUxykkY0wnS0%2flNY77KMj6E2KkHjMGfO1UsY%2futgUCHLpKwliR6IqU265LbeFKy9VSRcWwe28VRXFZtZeivYEF%2fkWvHL8p9rGpTeVZVtfhHgPreAJCGtI5uuC%2fEJYpbOb%2bw9GBZW%2bQw9I5VVIwXnX0gKm2m30%2b38AMdQe3UbAgAA&web=0

database("fabianm-java-perf-results").Results
| where TIMESTAMP > ago(5d)
| where Operation == "ReadLatency"
| where TIMESTAMP > datetime(2024-01-05 18:30:00.0000000)
| where BranchName startswith "origin:users/fabianm"
| where TestVariationName in ("Read_WithMetricsOnly_NoJavaAgent")
| extend Suffix=strcat(iif(BranchName=="origin:users/fabianm/mainclone", "Baseline", "This-PR"), ":", CommitId)
| extend Name=strcat(Suffix, ":", TestVariationName)
| summarize avg(SuccessRate) by bin(TIMESTAMP, 1h), Name
| render timechart

Tracing only (metrics disabled) with Java agent - but Reactor instrumentation disabled

About 3 percent improvement with this PR. When looking at CPU profiles I identified one issue in the AppInsights Java-Agent. It looks like a fix will be available with the next version of the Java Agent.

CPU Profile showing potential improvement in the Java Agent

Throughput improvement with this PR

https://fabianm-perf-results.westeurope.kusto.windows.net/fabianm-java-perf-results?query=H4sIAAAAAAAEAG1R3UsCQRB%2fF%2fwfhn06wcvTCkK4oKIHI01M6lHGvbm7CXcvdvf8iP74xlNKyH36MTu%2fj5nJMOASPUUqxyWjNfEHrjH%2bJJfHjny9Cl51LmYH1G59w6YkRzAfjR9f53fjKdwCFlV0nXX%2bPl%2bEjYErC2kKakaYPWMgq3fqvEAmv4ENRYNkcBUn%2fTi5hv7N8DIZJslFcngn8vcOrS4naAh8QBf8hkMJqnJcsB3WnpzvHYc5NSQf3tBxE6whs4WoSbd4F4G5Q822aPCTrOCuIBsWk0oadKjcyPrgaiO1RkA1eWgrY2XwWuc5b1Np0Bgi5jz6i5imZ4P1DLLVq8qS6oK6lwOs%2bIDnJft4OlMdwUMpPFTGcBhlp4aN8NHuYH5s%2fjdkw%2fK1MVL8IsB1IQStyfuZLL0Dyx0s2Ua%2f1%2bhC34jznrpnOnEjB%2fvj6FJW3W61Wz8mEzDrMgIAAA%3d%3d&web=0

database("fabianm-java-perf-results").Results
| where TIMESTAMP > ago(5d)
| where Operation == "ReadLatency"
| where TIMESTAMP > datetime(2024-01-05 18:30:00.0000000)
| where BranchName startswith "origin:users/fabianm"
| where TestVariationName in ("Read_WithTracing_WithJavaAgent_NoReactorInstrumentation")
| extend Suffix=strcat(iif(BranchName=="origin:users/fabianm/mainclone", "Baseline", "This-PR"), ":", CommitId)
| extend Name=strcat(Suffix, ":", TestVariationName)
| summarize avg(SuccessRate) by bin(TIMESTAMP, 1m), Name
| render timechart

All SDK Contribution checklist:

• The pull request does not introduce [breaking changes]
• CHANGELOG is updated for new features, bug fixes or other significant changes.
• I have read the contribution guidelines.

General Guidelines and Best Practices

• Title of the pull request is clear and informative.
• There are a small number of commits, each of which have an informative message. This means that previously merged commits do not appear in the history of the PR. For more information on cleaning up the commits in your PR, see this page.

Testing Guidelines

• Pull request includes test coverage for the included changes.

[azure-sdk]

API change check

API changes are not detected in this pull request.

[FabianMeiswinkel]

/azp run java - cosmos - tests

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[FabianMeiswinkel]

/azp run java - cosmos - tests

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[kushagraThapar]

LGTM, thanks @FabianMeiswinkel

[xinlian12]

LGTM, thanks