Latches and Resources (w)

img00100
Monitors latch/resource activity. Latches are usually held for just a few CPU cycles. Anything specific is already old. If the data persists across multiple intervals someone is possibly blocked and holding a latch.



Latch Actvity

ProTop RT Label Alertable Metric* Description
Latch N/A Latch code
Type latchType Latch type, usually "spin"
Hlder holder User Number of the last user holding the latch
QHold QHolder User number(s) of user(s) waiting for the latch
Requests latchReq Number of requests for that latch
Waits latchWts Number of times the process fails to get a latch because of excess spin
Lock% latchLkPct Percent of time the latch is locked


Resource Waits

ProTop RT Label Alertable Metric* Description
Resource N/A Resource name
Requests N/A Number of requests for that resource
Waits latchWts Number of times the process fails to get a latch because of excess spin
Lock% N/A Percent of time the resource is locked


TXE Locks & Waits

ProTop RT Label Alertable Metric* Description
TXE Type N/A Transaction type
Locks N/A The number of locks a transaction holds
Waits latchLkPct Number of times the process fails to get a latch because of excess spin
Lock% N/A Percent of time the transaction holds locks
ConLk% Not in use
ConLk Not in use
ConWt Not in use

*Requires data collector "LatchActivity" in pt3agent.cfg. With the commercial version of ProTop, you can configure alerts using the metric names in this column. For a full list of metrics, see the Alertable Metrics section.