[SERVER-49966] Examine performance bottlenecking of VersionAPI ServerStatus Metrics Created: 28/Jul/20  Updated: 27/Oct/23  Resolved: 08/Sep/20

Status: Closed
Project: Core Server
Component/s: Replication
Affects Version/s: None
Fix Version/s: 4.7.0

Type: Task Priority: Major - P3
Reporter: Pridhvi Vegesna (Inactive) Assignee: Xuerui Fa
Resolution: Works as Designed Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Issue Links:
Depends
depends on SERVER-49965 Add apiVersion metrics to serverStatu... Closed
Sprint: Repl 2020-09-07, Repl 2020-09-21
Participants:

 Description   

As part of changes in SERVER-49379, we acquire a mutex in every instance of execCommandDatabase in service_entry_point_common.cpp and runCommand in strategy.cpp  (within the addTimestamp function of APIVersionMetrics that is invoked by APIVersionMetrics::update) We also acquire this same mutex in the run method of ServerStatus cmd (generateSection method of APIVersionMetrics).

Examine the performance impact of SERVER-49379



 Comments   
Comment by Xuerui Fa [ 08/Sep/20 ]

It seems there was no noticeable regression after the metrics tickets were committed. I ran sys-perf patches with an increased workload for the InsertRemove task for before and after the commits, and it seems like the performance was similar in both cases. Closing this ticket!

Comment by A. Jesse Jiryu Davis [ 20/Aug/20 ]

Assigned to xuerui.fa since you're developing the metrics. Now that metrics are merged, ask David Daly for advice. If API version metrics do cause performance issues, when and where would we expect to see them in our perf testing?

Generated at Thu Feb 08 05:21:20 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.