| 9def8e3995b1034d943a0fd22b6e512bfacdab77 |
|
17-Apr-2018 |
Yangster-mac <yanglu@google.com> |
Reduce statsd log data size.
1. Hash the strings in metric dimensions.
2. Optimize the timestamp encoding in bucket.
Use bucket num for full bucket and millis for
partial bucket.
3. Encode the dimension path per metric and avoid
deduping it across dimensons.
Test: statsd test
Change-Id: I18f69654de85edb21a9c835c73edead756295e05
BUG: b/77813755
/frameworks/base/cmds/statsd/src/FieldValue.h
|
| e06cfd777a501eca17f8dd796b32ad61002c5875 |
|
11-Mar-2018 |
Yangster-mac <yanglu@google.com> |
Support slicing by chain.
BUG: b/73975181
Test: statsd test
Change-Id: I913ae0f68ff21ed0703bb5da9c60d3eaa3bf5981
/frameworks/base/cmds/statsd/src/FieldValue.h
|
| 13fb7e4eeaf7aee408821afe7ee55a5167e49e59 |
|
08-Mar-2018 |
Yangster <yanglu@google.com> |
Statsd cpu optimizaton
- cache the changed dimensions in condition tracker.
- avoid query condition wizard when unnecessary.
- avoid copy dimension keys in condition key generation.
Test: statsd tests.
BUG: b/73959649
Change-Id: I17d68e2a82643de3f421309841e75f84c6fd8f43
/frameworks/base/cmds/statsd/src/FieldValue.h
|
| 580ea321b16c71ddec515553761d6f37359bd3a0 |
|
26-Feb-2018 |
Yao Chen <yaochen@google.com> |
Add StateTracker.
StateTracker is a special condition tracker that's based on a state atom.
State atoms are annotated in atoms.proto.
The rules for StateTracker:
1. must not have "stop". must have "dimension"
2. must be based on a state atom.
3. it must have the all primary fields and the exclusive state field in its dimension.
For example UidProcessStateTracker, will have output dimension {uid, state}.
Test: unit tests added.
Change-Id: I6b77e58e9fabe61f7326daf929577d8b2cfbf27b
/frameworks/base/cmds/statsd/src/FieldValue.h
|
| 539288806fe5ae3733ef7883eb8ec01cce293988 |
|
26-Feb-2018 |
Yangster-mac <yanglu@google.com> |
Duration tracker optimization.
* Avoid querying sliced condition for stop/stopAll events for duration metric.
* Avoid extracting the internal dimension key when it is identical to the what dimension.
Test: statsd test
Change-Id: I664e8d3b1a68960d05c9ce4789caefb60b1ab502
/frameworks/base/cmds/statsd/src/FieldValue.h
|
| f520492279ce5c0a029c36b5251031a66f1413fb |
|
23-Feb-2018 |
Yangster-mac <yanglu@google.com> |
Avoid creating temporary objects in FiterValue().
It reduces the cpu time from 1000ns to 750ns
Test: statsd test.
Change-Id: Ifa7e98e3368f8d55f85c7b09d05a6c416482981d
/frameworks/base/cmds/statsd/src/FieldValue.h
|
| 4c959cb99eb7e71e5417a61f5429c5fa0073e826 |
|
13-Feb-2018 |
Yao Chen <yaochen@google.com> |
Remove unused variables in statsd, and make more warnings show.
Test: statsd_test
Change-Id: I2c7b674cb615f22c5de90c2de5f2d58108ab2e7f
/frameworks/base/cmds/statsd/src/FieldValue.h
|
| 8a8d16ceea1e5b7a2f8c41e17b5d993035f50f5d |
|
08-Feb-2018 |
Yao Chen <yaochen@google.com> |
Statsd CPU optimization.
The key change is to revamp how we parse/store/match a log event, especially how we match repeated
field and attribution nodes, and how we construct dimensions and compare them.
+ We use a integer to encode the field of a log element. And also encode the FieldMatcher into an
integer and a bit mask. The log matching becomes 2 integer operations.
+ Dimension is stored as encoded field and value pair. Checking if 2 dimensions are equal is then
becoming checking if the underlying integers are equal. The integers are stored contiguously
in memory, so it's much faster than previous tree structure.
Start review from FieldValue.h
Test: statsd_test + new unit tests
Bug: 72659059
Change-Id: Iec8daeacdd3f39ab297c10ab9cd7b710a9c42e86
/frameworks/base/cmds/statsd/src/FieldValue.h
|