Improve Tracing#8
Conversation
| import org.apache.cassandra.distributed.api.IMessageFilters; | ||
| import org.apache.cassandra.distributed.api.Row; | ||
| import org.apache.cassandra.distributed.api.SimpleQueryResult; | ||
| import org.apache.cassandra.distributed.shared.Metrics; |
There was a problem hiding this comment.
nit: looks like an unnecessary import
| return added; | ||
| } | ||
|
|
||
| if (++bucketSeen < 0) |
There was a problem hiding this comment.
Do we need to increment once again here given we have just incremented on L648?
| if (intersects != null && (participants == null || !intersects.intersects(participants))) | ||
| return false; | ||
|
|
||
| return chance >= 1.0f || ThreadLocalRandom.current().nextFloat() > chance; |
There was a problem hiding this comment.
Should this be ThreadLocalRandom.current().nextFloat() <= chance. For example, if we say chance is 0.95 (i.e. there is 95% chance of hitting this), we should return matches for 1.0 (and above), but also anything between 0 and 0.95
| { | ||
| if (keys != 0) out.append(','); | ||
| if (ranges == ALL_KINDS) out.append("R*"); | ||
| else TinyEnumSet.append(keys, Txn.Kind::forOrdinal, k -> "R" + k.shortName(), out); |
There was a problem hiding this comment.
Should this be TinyEnumSet.append(ranges?
|
|
||
| cur = new TxnEvents(); | ||
| if (newBucketSubSize < 0) | ||
| cur.bucketSubSize = newBucketSize; |
There was a problem hiding this comment.
Should we always set bucket sub size to new bucket size if newBucketjsubSize < 0, and not only when cur == null?
In other words, move this if outside of cur == null if statement?
There was a problem hiding this comment.
-1 means the user hasn't specified a value, but if we aren't inserting we don't need to update it (since there will already be some value). this is just for initialising to a default value
belliottsmith
force-pushed
the
remote-vtables
branch
from
12d3157 to
0517436
Compare
belliottsmith
force-pushed
the
improve-tracing
branch
from
fdaaba6 to
80322a9
Compare
belliottsmith
force-pushed
the
remote-vtables
branch
4 times, most recently
from
b61f3cb to
5888fc1
Compare
…irtual tables from any node in the cluster patch by Benedict; reviewed by Alex Petrov for CASSANDRA-20900
belliottsmith
force-pushed
the
improve-tracing
branch
6 times, most recently
from
baa8f18 to
0ec190b
Compare
|
good catches! |
- Introduce pattern tracing, that can intercept failed or new coordinations matching various filters
- Support additional tracing event collection modes (SAMPLE and RING)
patch by Benedict; reviewed by Alex Petrov for CASSANDRA-20911
belliottsmith
force-pushed
the
improve-tracing
branch
from
0ec190b to
c2dc7f7
Compare
2 participants
Opening up a separate PR since original tracing PR contains several other commits.