The following counters are available when monitoring Oracle 10G from LoadRunner. They can be captured at the "system" level using V$SYSSTAT, or captured by user session using V$SESSTAT. A friendly DBA indicated that those shown underlined would be most useful

active txn count during cleanout
application wait time
auto extends on undo tablespace
background checkpoints completed
bbackground timeouts
branch node splits
buffer is not pinned count
buffer is pinned count
bytes received via SQL*Net from client
bytes received via SQL*Net from dblink
bytes sent via SQL*Net to client
bytes sent via SQL*Net to dblink
Cached Commit SCN referenced
calls to get snapshot scn: kcmgss
calls to kcmgas
calls to kcmgcs
calls to kcmgrs
CCursor + sql area evicted
change write time
cleanout - number of ktugct calls
cleanouts and rollbacks - consistent read gets
cleanouts only - consistent read gets
cluster key scan block gets
cluster key scans
cluster wait time
cold recycle reads
commit batch performed
commit batch requested
commit batch/immediate performed
commit batch/immediate requested
commit cleanout failures: block lost
commit cleanout failures: buffer being written
commit cleanout failures: callback failure
commit cleanout failures: cannot pin
commit cleanout failures: hot backup in progress
commit cleanout failures: write disabled
commit cleanouts
commit cleanouts successfully completed
commit immediate performed
commit immediate requested
commit nowait performed
commit nowait requested
Commit SCN cached
commit txn count during cleanout
commit wait performed
commit wait requested
commit wait/nowait performed
commit wait/nowait requested
concurrency wait time
consistent changes
consistent gets
consistent gets - examination
consistent gets direct
consistent gets from cache
CPU used by this session
CPU used when call started
CR blocks created
current blocks converted for CR
cursor authentications
data blocks consistent reads - undo records applied
db block changes
db block gets
db block gets direct
db block gets from cache
DB time
DBWR checkpoint buffers written
DBWR checkpoints
DBWR fusion writes
DBWR lru scans
DBWR make free requests
DBWR object drop buffers written
DBWR parallel query checkpoint buffers written
DBWR revisited being-written buffer
DBWR tablespace checkpoint buffers written
DBWR thread checkpoint buffers written
DBWR transaction table writes
DBWR undo block writes
DDL statements parallelized
deferred (CURRENT) block cleanout applications
deferred CUR cleanouts (index blocks)
DFO trees parallelized
dirty buffers inspected
DML statements parallelized
doubling up with imu segment
drop segment calls in space pressure
enqueue conversions
enqueue deadlocks
enqueue releases
enqueue requests
enqueue timeouts
enqueue waits
exchange deadlocks
execute count
failed probes on index block reclamation
flashback log writes
frame signature mismatch
free buffer inspected
free buffer requested
gc blocks corrupt
gc blocks lost
gc claim blocks lost
gc CPU used by this session
gc cr block build time
gc cr block flush time
gc cr block receive time
gc cr block send time
gc cr blocks received
gc cr blocks served
gc current block flush time
gc current block pin time
gc current block receive time
gc current block send time
gc current blocks received
gc current blocks served
gc local grants
gc remote grants
gcs messages sent
ges messages sent
global enqueue CPU used by this session
global enqueue get time
global enqueue gets async
global enqueue gets sync
global enqueue releases
global undo segment hints helped
global undo segment hints were stale
heap block compress
hot buffers moved to head of LRU
immediate (CR) block cleanout applications
immediate (CURRENT) block cleanout applications
immediate CR cleanouts (index blocks)
IMU bind flushes
IMU commits
IMU contention
IMU CR rollbacks
IMU- failed to get a private strand
IMU Flushes
IMU ktichg flush
IMU mbu flush
IMU pool not allocated
IMU recursive-transaction flush
IMU Redo allocation size
IMU undo allocation size
IMU undo retention flush
index crx upgrade (found)
index crx upgrade (positioned)
index crx upgrade (prefetch)
index fast full scans (direct read)
index fast full scans (full)
index fast full scans (rowid ranges)
index fetch by key
index scans kdiixs1
instance recovery database freeze count
IPC CPU used by this session
java call heap collected bytes
java call heap collected count
java call heap gc count
java call heap live object count
java call heap live object count max
java call heap live size
java call heap live size max
java call heap object count
java call heap object count max
java call heap total size
java call heap total size max
java call heap used size
java call heap used size max
java session heap collected bytes
java session heap collected count
java session heap gc count
java session heap live object count
java session heap live object count max
java session heap live size
java session heap live size max
java session heap object count
java session heap object count max
java session heap used size
java session heap used size max
leaf node 90-10 splits
leaf node splits
lob reads
lob writes
lob writes unaligned
local undo segment hints helped
local undo segment hints were stale
logons cumulative
logons current
messages received
messages sent
Misses for writing mapping
native hash arithmetic execute
native hash arithmetic fail
no buffer to keep pinned count
no work - consistent read gets
number of map misses
number of map operations
opened cursors cumulative
opened cursors current
OTC commit optimization attempts
OTC commit optimization failure - setup
OTC commit optimization hits
Parallel operations downgraded 1 to 25 pct
Parallel operations downgraded 25 to 50 pct
Parallel operations downgraded 50 to 75 pct
Parallel operations downgraded to serial
Parallel operations not downgraded
parse count (failures)
parse count (hard)
parse count (total)
parse time cpu
parse time elapsed
physical read bytes
physical read IO requests
physical read total bytes
physical read total IO requests
physical read total multi block requests
physical reads
physical reads cache
physical reads cache prefetch
physical reads direct
physical reads direct (lob)
physical reads direct temporary tablespace
physical reads for flashback new
physical reads prefetch warmup
physical reads retry corrupt
physical write bytes
physical write IO requests
physical write total bytes
physical write total IO requests
physical write total multi block requests
physical writes
physical writes direct
physical writes direct (lob)
physical writes direct temporary tablespace
physical writes from cache
physical writes non checkpoint
pinned buffers inspected
prefetch clients - 16k
prefetch clients - 2k
prefetch clients - 32k
prefetch clients - 4k
prefetch clients - 8k
prefetch clients - default
prefetch clients - keep
prefetch clients - recycle
prefetch warmup blocks aged out before use
prefetch warmup blocks flushed out before use
prefetched blocks aged out before use
process last non-idle time
PX local messages recv'd
PX local messages sent
PX remote messages recv'd
PX remote messages sent
queries parallelized
queue flush
queue ocp pages
queue position update
queue qno pages
queue single row
queue splits
queue update without cp update
recovery array read time
recovery array reads
recovery blocks read
recursive aborts on index block reclamation
recursive calls
recursive cpu usage
redo blocks read for recovery
redo blocks written
redo buffer allocation retries
redo entries
redo log space requests
redo log space wait time
redo log switch interrupts
redo ordering marks
redo size
redo subscn max counts
redo synch time
redo synch writes
redo wastage
redo write time
redo writer latching time
redo writes
rollback changes - undo records applied
rollbacks only - consistent read gets
RowCR - resume
RowCR - row contention
RowCR attempts
RowCR hits
rows fetched via callback
serializable aborts
session connect time
session cursor cache count
session cursor cache hits
session logical reads
session pga memory
session pga memory max
session stored procedure space
session uga memory
session uga memory max
shared hash latch upgrades - no wait
shared hash latch upgrades - wait
SMON posted for dropping temp segment
SMON posted for instance recovery
SMON posted for supplemental logging cleanup
SMON posted for txn recovery for other instances
SMON posted for undo segment recovery
SMON posted for undo segment shrink
sorts (disk)
sorts (memory)
sorts (rows)
space was found by tune down
space was not found by tune down
sql area evicted
sql area purged
SQL*Net roundtrips to/from client
SQL*Net roundtrips to/from dblink
steps of tune down ret. in space pressure
summed dirty queue length
switch current to new buffer
table fetch by rowid
table fetch continued row
table lookup prefetch client count
table scan blocks gotten
table scan rows gotten
table scans (cache partitions)
table scans (direct read)
table scans (long tables)
table scans (rowid ranges)
table scans (short tables)
total number of slots
total number of times SMON posted
total number of undo segments dropped
transaction lock background get time
transaction lock background gets
transaction lock foreground requests
transaction lock foreground wait time
transaction rollbacks
transaction tables consistent read rollbacks
transaction tables consistent reads - undo records applied
tune down retentions in space pressure
undo change vector size
undo segment header was pinned
user calls
user commits
user I/O wait time
user rollbacks
workarea executions - multipass
workarea executions - onepass
workarea executions - optimal
workarea memory allocated
Workload Capture: dbtime
Workload Capture: errors
Workload Capture: size (in bytes) of recording
Workload Capture: unreplayable user calls
Workload Capture: unsupported user calls
Workload Capture: user calls
Workload Capture: user calls flushed
Workload Capture: user logins
Workload Capture: user txns
write clones created for recovery
write clones created in background
write clones created in foreground